CN115062605A - Service problem attribution method and device - Google Patents
Service problem attribution method and device Download PDFInfo
- Publication number
- CN115062605A CN115062605A CN202210763456.8A CN202210763456A CN115062605A CN 115062605 A CN115062605 A CN 115062605A CN 202210763456 A CN202210763456 A CN 202210763456A CN 115062605 A CN115062605 A CN 115062605A
- Authority
- CN
- China
- Prior art keywords
- text data
- analyzed
- service
- service problem
- classification result
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 106
- 239000013598 vector Substances 0.000 claims description 421
- 238000004458 analytical method Methods 0.000 claims description 73
- 238000012549 training Methods 0.000 claims description 68
- 238000013145 classification model Methods 0.000 claims description 10
- 230000009286 beneficial effect Effects 0.000 abstract description 4
- 235000013305 food Nutrition 0.000 description 97
- 238000011156 evaluation Methods 0.000 description 43
- 235000012054 meals Nutrition 0.000 description 19
- 238000012545 processing Methods 0.000 description 17
- 238000010586 diagram Methods 0.000 description 14
- 244000203593 Piper nigrum Species 0.000 description 11
- 235000008184 Piper nigrum Nutrition 0.000 description 11
- 235000013614 black pepper Nutrition 0.000 description 11
- 241001674044 Blattodea Species 0.000 description 10
- 235000013399 edible fruits Nutrition 0.000 description 9
- 241000251468 Actinopterygii Species 0.000 description 8
- 235000015278 beef Nutrition 0.000 description 8
- 230000006866 deterioration Effects 0.000 description 8
- 235000021108 sauerkraut Nutrition 0.000 description 8
- 235000013527 bean curd Nutrition 0.000 description 7
- 238000004590 computer program Methods 0.000 description 7
- 230000007246 mechanism Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 239000000843 powder Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 208000019331 Foodborne disease Diseases 0.000 description 3
- 241000238631 Hexapoda Species 0.000 description 3
- 239000004615 ingredient Substances 0.000 description 3
- 239000002075 main ingredient Substances 0.000 description 3
- 241000237858 Gastropoda Species 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 2
- 235000007164 Oryza sativa Nutrition 0.000 description 2
- 241000607479 Yersinia pestis Species 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 235000013361 beverage Nutrition 0.000 description 2
- 230000002457 bidirectional effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000001276 controlling effect Effects 0.000 description 2
- 230000008451 emotion Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 235000021022 fresh fruits Nutrition 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 230000002829 reductive effect Effects 0.000 description 2
- 235000009566 rice Nutrition 0.000 description 2
- 235000013311 vegetables Nutrition 0.000 description 2
- 235000010149 Brassica rapa subsp chinensis Nutrition 0.000 description 1
- 235000000536 Brassica rapa subsp pekinensis Nutrition 0.000 description 1
- 241000499436 Brassica rapa subsp. pekinensis Species 0.000 description 1
- 241000282320 Panthera leo Species 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000013065 commercial product Substances 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 235000005686 eating Nutrition 0.000 description 1
- 235000006694 eating habits Nutrition 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000000873 masking effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the application provides a service problem attribution method and a service problem attribution device. And analyzing the service problem attribution model according to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed to obtain a service problem classification result aiming at the text data to be analyzed. The service problem attribution model provided by the method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information, and the service knowledge graph can acquire the attribute information of the service information in the text data to be analyzed, so that the comprehension degree of the language expression logic of the text data to be analyzed is improved. The service knowledge graph is beneficial to improving the accuracy of the service problem attribution model for determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model for matching the text data to be analyzed with the service problem classification result is improved.
Description
Technical Field
The application relates to the technical field of computers, in particular to a service problem attribution method, a service problem attribution device, electronic equipment and a computer storage medium, another service problem attribution method, device, electronic equipment and a computer storage medium, and a service problem attribution model training method, device, electronic equipment and a computer storage medium.
Background
At present, various services are provided for a user by a merchant online platform, and after the user uses the services, the online platform evaluates the services. In order to quickly analyze problems in the evaluation information provided by the user and aiming at the service information, the online platform analyzes the evaluation information provided by the user by using an analysis model and identifies the problem type corresponding to the evaluation information.
In the prior art, the method for analyzing the problem type of the merchant service information generally has the problems of wrong identification result and low identification accuracy. Therefore, how to improve the accuracy of analyzing the problem types in the evaluation information is a problem to be solved.
Disclosure of Invention
The embodiment of the application provides a service problem attribution method, so that the accuracy of analyzing and evaluating the problem types in information is improved. The application also relates to a service problem attribution device, electronic equipment and a computer storage medium, another service problem attribution method, another service problem attribution device, another service problem attribution electronic equipment and a computer storage medium, and a service problem attribution model training method, a service problem attribution device, another service problem attribution electronic equipment and a computer storage medium.
The embodiment of the application provides a service problem attribution method, which comprises the following steps: acquiring text data to be analyzed aiming at service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, and obtaining a first embedded vector aiming at the text data to be analyzed, wherein the first embedded vector comprises a text data to be analyzed embedded vector and a service knowledge graph embedded vector related to the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a first embedded vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model to obtain text data embedding vectors to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed; and obtaining a first embedded vector aiming at the text data to be analyzed according to the text data embedded vector to be analyzed and the service knowledge map embedded vector.
Optionally, the obtaining, according to the first embedded vector, a service problem classification result corresponding to the text data to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the first embedded vector; and according to the characteristic information of the text data to be analyzed, inquiring a target service problem classification result matched with the characteristic information of the text data to be analyzed in a service problem classification list of the service problem attribution model to serve as a service problem classification result corresponding to the text data to be analyzed.
Optionally, the querying, according to the feature information of the text data to be analyzed, a target service problem classification result matched with the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model, as a service problem classification result corresponding to the text data to be analyzed, includes: acquiring characteristic information corresponding to a plurality of candidate service problem classification results in a service problem classification list of the service problem attribution model; and comparing the feature information of the text data to be analyzed with the feature information respectively corresponding to the candidate service problem classification results, and determining the candidate service problem classification result containing the feature information of the text data to be analyzed as a target service problem classification result matched with the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed.
Optionally, the obtaining, according to the first embedded vector, a service problem classification result corresponding to the text data to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the first embedded vector; according to the characteristic information of the text data to be analyzed, at least one service problem classification result matched with the characteristic information of the text data to be analyzed is inquired in a service problem classification list of the service problem attribution model; and taking the at least one service problem classification result as a service problem classification result corresponding to the text data to be analyzed.
Optionally, the obtaining, according to the first embedded vector, a service problem classification result corresponding to the text data to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the first embedded vector; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
Optionally, the service problem attribution model comprises a large-class text data feature analysis model, and the large-class text data feature analysis model is used for analyzing feature information of text data corresponding to a large-class result of a service problem; the querying, according to the feature information of the text data to be analyzed, a service problem large classification result associated with the feature information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model, as a target service problem large classification result of the text data to be analyzed, includes: according to the large classification text data feature analysis model, first feature information of text data used for analyzing the large classification result of the service problem in the text data to be analyzed is obtained; acquiring first candidate characteristic information of the text data corresponding to each candidate service problem large classification result in the service problem large classification list; and comparing the first characteristic information in the text data to be analyzed with the first candidate characteristic information corresponding to each candidate service problem large classification result to obtain a target service problem large classification result corresponding to the text data to be analyzed.
Optionally, on the basis of determining the target service problem large classification result of the text data to be analyzed, querying a target service problem small classification result associated with the feature information of the text data to be analyzed in the service problem small classification list of the target service problem large classification result includes: acquiring second characteristic information of the text data used for analyzing the small classification result of the service problem in the text data to be analyzed; acquiring second candidate characteristic information of the text data corresponding to each candidate service problem small classification result in the service problem small classification list; and comparing the second characteristic information in the text data to be analyzed with the second candidate characteristic information corresponding to each candidate service problem small classification result to obtain a target service problem small classification result corresponding to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into the service problem attribution model to obtain the text data embedding vector to be analyzed includes: inputting the text data to be analyzed into the service problem cause model, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, wherein each text unit embeds a vector in a corresponding paragraph in the text data to be analyzed, and each text unit embeds a vector in a corresponding position in the text data to be analyzed; embedding a text embedding vector corresponding to each text unit in the text data to be analyzed, and embedding a vector corresponding to each text unit in the text data to be analyzed at a corresponding position in the text data to be analyzed, and processing to obtain the text data embedding vector to be analyzed.
Optionally, the querying a service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed includes: acquiring a service information embedded vector in the text data to be analyzed according to the text data embedded vector to be analyzed; according to the service information embedding vector in the text data to be analyzed, inquiring a service knowledge graph aiming at the service information; acquiring a text embedding vector corresponding to each text unit in a service knowledge graph aiming at the service information, wherein each text unit is embedded with a vector in a corresponding paragraph in the service knowledge graph, and each text unit is embedded with a vector in a corresponding position in the service knowledge graph; embedding a text corresponding to each text unit in a service knowledge graph aiming at the service information into a vector, embedding each text unit into a corresponding paragraph in the service knowledge graph, and embedding each text unit into a vector at a corresponding position in the service knowledge graph for processing to obtain the service knowledge graph embedding vector aiming at the service information.
Optionally, the service information is food service information provided by the merchant to the user for food service; the analysis text data aiming at the service information is evaluation information of a user aiming at the food service information; the obtaining of the text data to be analyzed for the service information includes: and acquiring text data to be analyzed aiming at the food service information, which is sent by a user side.
Optionally, the service problem attribution model is specifically configured to obtain a service problem classification result for the text data to be analyzed according to the text data to be analyzed, pinyin information corresponding to the text data to be analyzed, and a service knowledge graph related to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, and obtaining a second embedded vector aiming at the text data to be analyzed, wherein the second embedded vector comprises a pinyin embedded vector of the text data to be analyzed and a pinyin embedded vector of a service knowledge graph related to the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the second embedded vector.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a second embedded vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise pinyin embedding vectors corresponding to each text unit in the text data to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed, wherein the service knowledge map embedding vector comprises a pinyin embedding vector corresponding to each text unit in the service knowledge map; and obtaining a second embedded vector aiming at the text data to be analyzed according to the text data embedded vector to be analyzed and the service knowledge map embedded vector.
Optionally, the inputting the text data to be analyzed into the service problem attribution model to obtain the text data embedding vector to be analyzed includes: inputting the text data to be analyzed into the service problem attribution model, obtaining text embedded vectors corresponding to each text unit in the text data to be analyzed, pinyin embedded vectors corresponding to each text unit, paragraph embedded vectors corresponding to each text unit in the text data to be analyzed, and embedded vectors corresponding to each text unit in the text data to be analyzed; embedding a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed, and processing to obtain the text data embedding vector to be analyzed.
Optionally, the querying a service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed includes: acquiring a service information embedded vector in the text data to be analyzed according to the text data embedded vector to be analyzed; according to the service information embedding vector in the text data to be analyzed, inquiring a service knowledge graph aiming at the service information; acquiring a text embedded vector corresponding to each text unit in a service knowledge graph aiming at the service information, a pinyin embedded vector corresponding to each text unit, a paragraph embedded vector corresponding to each text unit in the service knowledge graph, and a position embedded vector corresponding to each text unit in the service knowledge graph; embedding a text corresponding to each text unit in a service knowledge graph aiming at the service information into a vector, embedding a pinyin corresponding to each text unit into a vector, embedding each text unit into a corresponding paragraph in the service knowledge graph, and embedding each text unit into a vector at a corresponding position in the service knowledge graph for processing to obtain the service knowledge graph embedding vector aiming at the service information.
The embodiment of the present application further provides a service problem attribution method, including: acquiring text data to be analyzed aiming at service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise text embedding vectors corresponding to each text unit in the text data to be analyzed, pinyin embedding vectors corresponding to each text unit, paragraph embedding vectors corresponding to each text unit in the text data to be analyzed, and embedding vectors corresponding to each text unit in the position of the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain an embedded vector of the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, carrying out vectorization processing on the text data to be analyzed, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a vector embedded in a corresponding position of each text unit in the text data to be analyzed; embedding a vector into a text corresponding to each text unit in the text data to be analyzed, embedding a vector into a pinyin corresponding to each text unit, embedding a vector into a paragraph corresponding to each text unit in the text data to be analyzed, and embedding a vector into a position corresponding to each text unit in the text data to be analyzed for processing to obtain the text data embedding vector to be analyzed.
Optionally, the obtaining a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the embedded vector of the text data to be analyzed; and according to the characteristic information of the text data to be analyzed, inquiring a target service problem classification result matched with the characteristic information of the text data to be analyzed in a service problem classification list of the service problem attribution model to serve as a service problem classification result corresponding to the text data to be analyzed.
Optionally, the obtaining a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the text data embedding vector to be analyzed; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
Optionally, the service problem attribution model is specifically configured to obtain a service problem classification result for the text data to be analyzed according to the text data to be analyzed, pinyin information corresponding to the text data to be analyzed, and a service knowledge graph related to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, and obtaining a second embedded vector aiming at the text data to be analyzed, wherein the second embedded vector comprises a text data to be analyzed embedded vector and a service knowledge graph embedded vector related to the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the second embedded vector.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a second embedded vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise pinyin embedding vectors corresponding to each text unit in the text data to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed, wherein the service knowledge map embedding vector comprises pinyin embedding vectors corresponding to each text unit in the service knowledge map; and obtaining a second embedded vector aiming at the text data to be analyzed according to the embedded vector of the text data to be analyzed and the embedded vector of the service knowledge graph.
Optionally, the obtaining, according to the second embedded vector, a service problem classification result corresponding to the text data to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the second embedded vector; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model, and taking the service problem large classification result as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
The embodiment of the present application further provides a training method for a service problem attribution model, including: obtaining a pre-training model for analyzing feature information of text data; adjusting the pre-training model according to a text data sample aiming at service information and a service problem big classification result sample aiming at the text data sample to obtain a big classification text data feature analysis model, wherein the big classification text data feature analysis model is used for analyzing feature information of text data corresponding to the service problem big classification result; and adjusting the large-classification text data feature analysis model according to the text data sample aiming at the service information and the small-classification result sample aiming at the service problem of the text data sample to obtain a service problem attribution model for analyzing the classification result aiming at the service problem of the text data to be analyzed.
Optionally, the pre-training model analyzes feature information of the text data in the following manner: obtaining text data embedded vectors according to the text data, wherein the text data embedded vectors comprise text embedded vectors corresponding to each text unit in the text data, pinyin embedded vectors corresponding to each text unit, paragraph embedded vectors of each text unit in the text data, and position embedded vectors of each text unit in the text data; and analyzing the characteristic information of the text data according to the text data embedded vector.
Optionally, the method further includes: acquiring a service knowledge graph embedding vector associated with the text data according to the text data embedding vector, wherein the service knowledge graph embedding vector comprises a text embedding vector corresponding to each text unit in a service knowledge graph associated with the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph; the analyzing the feature information of the text data according to the text data embedding vector comprises: analyzing feature information of the text data according to the text data embedding vector and a service knowledge graph embedding vector associated with the text data.
Optionally, the adjusting the pre-training model according to the text data sample for the service information and the service problem large classification result sample for the text data sample to obtain a large classification text data feature analysis model includes: inputting the text data sample aiming at the service information into the pre-training model to obtain a first service problem big classification result of the text data sample aiming at the service information output by the pre-training model; and adjusting the service problem major classification parameters of the pre-training model according to the similarity between the first service problem major classification result and the service problem major classification result sample aiming at the text data sample to obtain the major classification text data feature analysis model.
Optionally, the inputting the text data sample for the service information into the pre-training model to obtain a first service problem large classification result of the text data sample for the service information output by the pre-training model includes: inputting the text data sample aiming at the service information into the pre-training model to obtain the characteristic information of the text data sample; and obtaining a first service problem big classification result of the text data sample aiming at the service information and output by the pre-training model according to the characteristic information of the text data sample.
Optionally, the adjusting the large-category text data feature analysis model according to the text data sample for the service information and the small service problem classification result sample for the text data sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed includes: inputting text data aiming at the service information into the large-class text data characteristic analysis model to obtain a second small service problem classification result aiming at the text data sample and output by the large-class text data characteristic analysis model; and adjusting the large-class text data feature analysis model according to the similarity degree between the second small service problem classification result and the small service problem classification result sample to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed.
Optionally, the method further includes: taking a text data sample aiming at service information and a service problem large-classification result sample which does not belong to the text data as a first negative sample pair, and adjusting the pre-training model to obtain a large-classification text data feature analysis model for analyzing feature information of the text data corresponding to the service problem large-classification result; and taking a text data sample aiming at the service information and a service problem small classification result sample which does not belong to the text data as a second negative sample pair, and adjusting the large classification text data feature analysis model to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed.
An embodiment of the present application further provides a service problem attribution device, including: a first obtaining unit configured to obtain text data to be analyzed for service information; the second obtaining unit is used for inputting the text data to be analyzed into a service problem attribution model and obtaining a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
An embodiment of the present application further provides a service problem attribution device, including: a third obtaining unit configured to obtain text data to be analyzed for the service information; a fourth obtaining unit, configured to input the text data to be analyzed into a service problem attribution model, and obtain a service problem classification result for the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.
The embodiment of the present application further provides a training device for a service problem attribution model, including: a pre-training model obtaining unit for obtaining a pre-training model for analyzing feature information of text data; a large-category text data feature analysis model obtaining unit, configured to adjust the pre-training model according to a text data sample for service information and a large-category service problem result sample for the text data sample to obtain a large-category text data feature analysis model, where the large-category text data feature analysis model is used to analyze feature information of text data corresponding to a large-category service problem result; and the service problem attribution model obtaining unit is used for adjusting the large-classification text data feature analysis model according to the text data sample aiming at the service information and the small-classification result sample aiming at the service problem of the text data sample to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed.
The embodiment of the application also provides an electronic device, which comprises a processor and a memory; the memory stores a computer program, and the processor executes the computer program and then executes the method.
An embodiment of the present application further provides a computer storage medium, in which a computer program is stored, and the computer program executes the method.
Compared with the prior art, the embodiment of the application has the following advantages:
the embodiment of the application provides a service problem attribution method, which comprises the following steps: acquiring text data to be analyzed aiming at service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
In one embodiment of the application, the service problem attribution model analyzes and obtains a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed. The service problem attribution model provided by the method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information, and the service knowledge graph can acquire the attribute information of the service information in the text data to be analyzed, so that the comprehension degree of the language expression logic of the text data to be analyzed is improved. The service knowledge graph is beneficial to improving the accuracy of the service problem attribution model for determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model for matching the text data to be analyzed with the service problem classification result is improved.
In an embodiment of the application, the service problem attribution model analyzes and obtains a service problem classification result for the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed. The service problem attribution model provided by the method combines the pinyin information of the text data to be analyzed when the text data to be analyzed is processed, and improves the accuracy of the language expression logic of the text data to be analyzed. The accuracy of the language expression logic of the text data to be analyzed is improved based on the text data to be analyzed, and the accuracy of the service problem attribution model for matching the text data to be analyzed with the service problem classification result is improved.
In one embodiment of the present application, there is provided a training method of a service problem attribution model, comprising: obtaining a pre-training model for analyzing feature information of text data; adjusting the pre-training model according to a text data sample aiming at service information and a service problem big classification result sample aiming at the text data sample to obtain a big classification text data feature analysis model, wherein the big classification text data feature analysis model is used for analyzing feature information of text data corresponding to the service problem big classification result; and adjusting the large-classification text data feature analysis model according to the text data sample aiming at the service information and the small-classification result sample aiming at the service problem of the text data sample to obtain a service problem attribution model for analyzing the classification result aiming at the service problem of the text data to be analyzed.
According to the method, a pre-training model is obtained, firstly, a text data sample and a large classification result sample aiming at the service problem of the text data sample to be analyzed are adopted, the pre-training model is adjusted, and a large classification text data characteristic analysis model is obtained. And then, adjusting the large-classification text data characteristic analysis model according to the text data sample and the small service problem classification result sample aiming at the text data sample to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed. Therefore, when the trained service problem attribution model obtains the service problem classification result of the text data to be analyzed, the service problem big classification result of the text data to be analyzed is obtained first, and on the basis of determining the service problem big classification result of the text data to be analyzed, the service problem small classification result of the text data to be analyzed is determined in a plurality of service problem small classification results corresponding to the target service problem big classification result. Therefore, the finally obtained service problem classification result of the text data to be analyzed comprises a service problem large classification result and a service problem small classification result.
Drawings
Fig. 1 is an application scenario diagram of a service problem attribution method according to an embodiment of the present application.
Fig. 2 is a schematic diagram of a service problem classification result for text information in a service problem attribution model according to an embodiment of the present application.
Fig. 3 is a display page of the merchant end physical examination center provided in the embodiment of the present application.
Fig. 4 is food safety risk detail information of merchant stores provided in the embodiment of the present application.
Fig. 5 is data information of the food safety risk detail information of fig. 4.
Fig. 6 shows the food safety risk problem level of the merchant in fig. 5 and a processing manner corresponding to the food safety risk level.
Fig. 7 is a food safety risk learning method established for a merchant according to an embodiment of the present application.
Fig. 8 is a flowchart of a service problem attribution method provided in the first embodiment of the present application.
Fig. 9 is a schematic diagram of a service problem attribution device provided in the second embodiment of the present application.
Fig. 10 is a schematic diagram of another service problem attribution method provided in the third embodiment of the present application.
Fig. 11 is a schematic view of another service problem attribution apparatus provided in the fourth embodiment of the present application.
Fig. 12 is a schematic diagram of a training method of a service problem attribution model provided in the fifth embodiment of the present application.
Fig. 13 is a schematic diagram of a training device for a service problem attribution model provided in a sixth embodiment of the present application.
Fig. 14 is a schematic view of an electronic device provided in a seventh embodiment of the present application.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is capable of implementation in many different ways than those herein set forth and of similar import by those skilled in the art without departing from the spirit and scope of this application, and thus this application is not limited to the specific implementations disclosed below.
The terminology used in the description herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The description used in this application and in the appended claims is for example: the terms "a," "an," "first," and "second," etc., are not intended to be limiting in number or order, but rather are used to distinguish one type of information from another.
The embodiment of the application firstly provides a service problem attribution method, a service problem attribution device, electronic equipment and a computer storage medium. The embodiment of the application also provides another service problem attribution method, device, electronic equipment and computer storage medium, and a service problem attribution model training method, device, electronic equipment and computer storage medium.
In order to show the service problem attribution method provided by the application more clearly, an application scenario of the service problem attribution method provided by the application is introduced first.
The service problem attribution method can be applied to food safety problem classification scenes. For example, an online-to-offline catering service system provides a user with a meal ordering service. Fig. 1 is an application scenario diagram of a service problem attribution method provided in the embodiment of the present application. The user orders food items through the user terminal 101, orders food items, and feeds back related data for the food items, such as evaluation information for the food items ordered, food safety insurance claim evaluation information, and evaluation data for the food items consumed. The evaluation information comprises complaint information of the safety problem of the user aiming at partial food. The server 102 receives the evaluation information for the food sent by the user, judges whether the evaluation information is complaint information for food safety problems of the food according to the evaluation information, determines a food safety problem type corresponding to the evaluation information if the evaluation information is complaint information for the food safety problems of the food, and sends the evaluation information and the food safety problem type for the evaluation information to a merchant providing the food.
The server 102 analyzes the food safety problem type corresponding to the evaluation information provided by the user through the service problem attribution model. Specifically, the obtained text data to be analyzed for the service information is input into a service problem attribution model, and a service problem classification result for the text data to be analyzed, which is output by the service problem attribution model, is obtained.
For example, the evaluation information of the user on the meal is input into the service problem attribution model, and the service problem attribution model analyzes the evaluation information to determine the classification result of the service problem corresponding to the evaluation information, namely the food safety attribution type corresponding to the meal. For example, the evaluation information of the user on the meal is: when the evaluation information is input into the service problem attribution model, the food safety attribution type corresponding to the food output by the service problem attribution model is visible mildewed deterioration. For another example, the evaluation information is: when the evaluation information is input into the service problem attribution model, the food safety attribution type corresponding to the food output by the service problem attribution model is the cockroach.
The service problem attribution model is used for determining a service problem classification result corresponding to evaluation information aiming at the service information provided by a user. Specifically, a large service problem classification result corresponding to the evaluation information is determined, and on the basis of determining the large service problem classification result, a small service problem classification result corresponding to the evaluation information is analyzed. And generating a service problem classification result aiming at the evaluation information according to the service problem large classification result and the service problem small classification result in the service problem large classification result. Taking the above evaluation information of "cockroach exists in meal" as an example, the large classification result of the service problem of the evaluation information is a foreign object, and the small classification result of the service problem is a cockroach, wherein the cockroach is a small classification result of the insect and mouse in the foreign object.
And after obtaining the food safety risk classification result corresponding to the evaluation information of the user, the server sends a food safety event existing in the food provided by the merchant and a problem classification result corresponding to the food safety event to the merchant. The merchant checks the food safety risk problem through a physical examination center of the merchant terminal.
Fig. 3 is a display page of the merchant terminal physical examination center according to the embodiment of the present application. In the page shown in fig. 3, the physical examination center at the merchant end shows the risk diagnosis condition of the merchant, the management and control processing records, the food safety credit score information of the merchant, and the learning special area information for the food safety problem. The page of fig. 3 shows the merchant the current food safety risk issues and the corresponding learning strategy that the merchant identifies and further corrects the current food safety issues. After the merchant clicks the view risk button on the page of fig. 3, the page shown in fig. 4 is entered.
Please refer to fig. 4, which is food safety risk detail information of merchant stores provided in the embodiment of the present application. In the page of fig. 4, the merchant may specifically connect the details of the food safety risk that exist in the current month, and the merchant may further improve the problem according to the details.
Fig. 5 is data information of the food safety risk details information of fig. 4. In fig. 5, the merchant can learn the food safety risk incidence that occurs in the month, and the number of food safety events.
Fig. 6 shows the food safety risk problem level of the merchant in fig. 5 and a processing manner corresponding to the food safety risk level.
Meanwhile, the server 102 provides the merchant terminal 103 used by the merchant with a learning course for dealing with the food safety problem type, so that the merchant can deal with the food safety problem in time and the food safety problem is prevented from occurring continuously.
Fig. 7 is a food safety risk learning method established for a merchant according to an embodiment of the present application. Specifically, fig. 7 illustrates a method for preventing and controlling insect and mouse pests, and specifically explains details of the method for preventing and controlling insect and mouse pests, corresponding regulatory requirements, and preventive measures.
In addition, the service problem attribution method provided by the embodiment of the application can also be applied to a scene of recommending food to the user. For example, a user logs in an online take-out platform, searches for related meal item keywords in a search box of the online take-out platform, a server of the online take-out platform obtains meal item information corresponding to the search keywords by adopting a corresponding recommendation algorithm according to the search keywords of the user, sends the obtained meal item information to a user side, and the user side displays the obtained meal item information.
When the server determines the corresponding recommended food according to the search keyword provided by the user, the factors referred by the recommendation algorithm of the server include multiple factors, such as eating habit information of the user, food characteristic information of the food searched by the user, geographical location information of the user, and food safety risk level corresponding to a merchant providing the food. The food safety risk level of the merchant can be obtained by adopting the service problem attribution model in the service problem attribution method provided by the application.
In addition, the service problem attribution method provided by the embodiment of the application can also be applied to merchant information recommended by a home page of the user side. For example, after the user logs in the online takeout platform, the merchant information displayed on the home page of the user side is obtained in advance by the server side through a recommendation algorithm. Wherein, the reference factors of the recommendation algorithm comprise: distance information between the geographical position of the merchant and the geographical position of the user, and the historical attention of the user on the type of the merchant, the type of food and the food safety risk level of the merchant. The food safety risk level of the merchant can be obtained by adopting the service problem attribution model in the service problem attribution method provided by the application.
The food safety risk level of a merchant can be obtained by the following method:
the user provides evaluation information aiming at the take-out products of the merchant, the service problem attribution model is adopted to analyze the evaluation information provided by the user, and the food safety problem in the evaluation information and the cause type causing the food safety problem are determined. That is, if the evaluation information provided by the user for the food eaten by the user includes a food safety problem, a specific safety problem classification result corresponding to the food safety problem is determined according to the service problem attribution model. And determining the food safety risk level of the merchant corresponding to the food according to the determined food safety problem classification result.
In the embodiment of the application, a service problem attribution model is adopted to analyze the classification result of the food safety problem contained in the user evaluation information. And sending the evaluation information aiming at the food and the food safety problem classification result corresponding to the food safety problem in the evaluation information to a merchant providing the food for the user. On one hand, the food safety awareness of the merchant is improved, and the severity of the food safety problem of the merchant is prompted. And on the other hand, specific food safety problems sent in the process of eating the food are prompted to the merchant. Third, merchants are provided with a way to address such food safety issues, avoiding them from later reoccurring. Wherein, if the types of the food safety problems exist in the merchant, certain compulsory change measures are provided for the merchant.
Based on the steps, the online takeout system can determine the food safety level of the commercial tenant, and when recommending the corresponding commercial tenant list to the user, whether the commercial tenant is recommended to the user is determined according to the food safety level of the commercial tenant.
The service problem attribution model mentioned in each application scene obtains a service problem classification result corresponding to the user evaluation information by analyzing the user evaluation information. The service problem attribution model is specifically analyzed by the following method:
first, a service problem attribution model determines a service problem classification result for text data to be analyzed according to the text data to be analyzed and a service knowledge graph associated with the text data to be analyzed.
The service knowledge graph refers to a priori knowledge graph of text data to be analyzed and is stored in the service problem attribution model in advance.
The service knowledge graph at least comprises the following types:
1) the service knowledge graph is a service knowledge graph attributed, that is, a knowledge graph of service information in the text data to be analyzed. For example, the text data to be analyzed is: the strong-smelling preserved bean curd is good in smell. The corresponding service knowledge graph is as follows: stinky tofu is a food whose odor attribute is stinky.
When the traditional analysis method is adopted to analyze the text data to be analyzed, the analysis result shows that the stinky tofu has the problem of deterioration.
However, in the embodiment of the present application, when the service problem attribution model analyzes the text data to be analyzed, namely "strong-smelling preserved bean curd is smelly", it may simultaneously combine the service knowledge map "strong-smelling preserved bean curd is a food, the smell attribute of the food is smelly", and analyze the service problem classification result corresponding to the text data to be analyzed. In this example, "strong-smelling fermented bean curd is good and" strong-smelling fermented bean curd is a food, and the information of the odor attribute of the food being bad "can be determined, and the question fed back in the text data to be analyzed does not belong to any attribution type in attribution standards of the attribution model of the service question, indicating that the question fed back by the text data to be analyzed does not belong to the service question.
The service problem attribution model carries out vectorization processing on input information, and output information is also represented by vectors. For example, a vector of 1 indicates that the service information of the text data to be analyzed corresponds to at least one attribution type in the attribution standard. A vector of 0 indicates that the service information of the text data to be analyzed does not correspond to any of the attribution types in the attribution standards.
For another example, the text data to be analyzed is: the powder is good in smell. The traditional analysis method can determine that the deterioration problem of the snail powder exists. However, in the embodiment of the application, when the service problem attribution model analyzes the text data to be analyzed, the service knowledge graph is combined. Here, the service knowledge graph corresponding to the text data to be analyzed is: the snail powder is a commercial product whose odor attribute is malodorous. Therefore, the service problem attribution model combines and analyzes the text data to be analyzed and the corresponding service knowledge map to obtain the product attribute of the lion powder, which indicates that the problem fed back in the text data to be analyzed does not belong to any attribution type in attribution standards of the service problem attribution model.
2) The service knowledge graph is an ingredient graph, for example, the text data to be analyzed is "how to have the black pepper woollen in the black pepper beef fillet", and the corresponding service knowledge graph is "the ingredient table of the black pepper beef fillet includes: black pepper and beef.
When the text data to be analyzed is analyzed by adopting a traditional analysis method, the analysis result is that foreign matters exist in the black pepper beef fillet. However, in the embodiment of the present application, when the service problem attribution model analyzes the user evaluation information "how to leave the black pepper in the black pepper beef fillet", the service knowledge map "the ingredient list of the black pepper beef fillet includes, in combination with the service knowledge map: black pepper and beef. Therefore, the black pepper in the black pepper beef fillet is one of the main ingredients of the dish. Therefore, it is stated that the problem fed back by the text data to be analyzed does not belong to any attribution type in the attribution standard in the service problem attribution model, which has no service problem.
3) The service knowledge graph is a menu, for example, the text data to be analyzed is 'which fish the sauerkraut is', and the corresponding service knowledge graph is 'the sauerkraut is a menu, and the main ingredients are sauerkraut and fish'.
When the traditional analysis method is adopted to analyze the text data to be analyzed, the analysis result is that the problem of the pickled Chinese cabbage fish is the problem of food deterioration. However, in the embodiment of the present application, when the service problem attribution model analyzes the text data to be analyzed, which kind of fish the sauerkraut is, the service problem attribution model combines with the corresponding service knowledge map, which the sauerkraut is a menu and the main ingredients are sauerkraut and fish. It was thus determined that sauerkraut fish is a dish made from common fish and sauerkraut, not the fish species name. Thus, it is stated that the problem fed back in the text data to be analyzed does not belong to any of the attribution types in the attribution standards in the service problem attribution model, which is free from service problems.
4) The service knowledge graph is a common sense graph. For example, the text data to be analyzed is "what you are", and the corresponding service knowledge graph is "what you are".
When the text data to be analyzed is analyzed by a conventional analysis method, the analysis result is a mouse. However, in the embodiment of the present application, when the service problem attribution model analyzes the text data to be analyzed, the service problem attribution model is "good for itself" in combination with the corresponding service knowledge graph. Therefore, the text data to be analyzed is determined by the fact that the user prompts the merchant, the service is relatively poor compared with the meal service provided by the merchant for the user before, but the service does not belong to any attribution type in attribution standards in a service problem attribution model, and the merchant is expected to be good and improve the current service quality.
The service problem attribution model is used for analyzing the language expression logic of the text data to be analyzed, and the accuracy of the language expression logic of the text data to be analyzed can be improved by combining the service knowledge map corresponding to the text data to be analyzed. Furthermore, after the language expression logic of the text data to be analyzed is obtained according to the text data to be analyzed and the service knowledge graph corresponding to the text data to be analyzed, the accuracy of determining the service classification result corresponding to the text data to be analyzed is also improved.
The service problem attribution model adopts the text data to be analyzed and the service knowledge graph associated with the text data to be analyzed, and the process of determining the service problem attribution result corresponding to the text data to be analyzed is as follows:
inputting text data to be analyzed into a service problem attribution model, and obtaining text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise text embedding vectors corresponding to each text unit in the text data to be analyzed, each text unit is embedded into a paragraph in the text data to be analyzed, and each text unit is embedded into a vector at a position in the text data to be analyzed. Wherein each text unit may be each word or each word. As in FIG. 1, E This is achieved by Text embedding vector, E, representing the correspondence of this word A A paragraph embedding vector, E, representing this word in the text data to be analyzed 1 Indicating that this word embeds a vector at the location in the text data to be analyzed.
Then, according to the text data embedding vector to be analyzed, obtaining a service knowledge graph embedding vector associated with the text data to be analyzed, wherein the service knowledge graph embedding vector comprises a text embedding vector corresponding to each text unit in a service knowledge graph, each text unit is embedded in a paragraph in the service knowledge graph, and each text unit is embedded in a position in the service knowledge graph.
And embedding the text data to be analyzed into a vector and a service knowledge graph embedding vector associated with the text data to be analyzed, and generating a first embedding vector aiming at the text data to be analyzed. The service problem attribution model analyzes the first embedded vector aiming at the text data to be analyzed, determines the characteristic information of the text data to be analyzed for analyzing the service problem classification result, and determines the service problem classification result corresponding to the text data to be analyzed according to the characteristic information of the text data to be analyzed.
The service problem analysis model determines a service problem classification result corresponding to the text data to be analyzed according to the characteristic information of the text data to be analyzed. Specifically, the large classification result of the service problem corresponding to the text data to be analyzed, such as the classification result of foreign matter, deterioration, food-borne disease, overdue food, immature food, etc. shown in fig. 2, is determined according to the feature information of the text data to be analyzed. The process of determining the result of the big classification of the target service problem is also called a fixed sharing layer. And then, after the sharing layer is fixed, selecting a target classification layer corresponding to the characteristic information of the text data to be analyzed, namely determining a target service problem small classification result of the text data to be analyzed. Such as the pins shown in fig. 2. And determining the small classification result of the target service problem, namely the classification result of the target service problem of the text data to be analyzed.
Secondly, the service problem attribution model determines a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information of the text data to be analyzed.
Specifically, text data to be analyzed is input into the service problem attribution model, text data embedding vectors to be analyzed are obtained, the text data embedding vectors to be analyzed include text embedding vectors corresponding to each text unit in the text data to be analyzed, pinyin embedding vectors corresponding to each text unit, vectors are embedded in paragraphs of each text unit in the text data to be analyzed, and vectors are embedded in positions of each text unit in the text data to be analyzed.
The text data embedding vector to be analyzed comprises a pinyin embedding vector corresponding to each text unit, so that the service problem attribution model converts the text data to be analyzed into the embedding vector, the phenomenon of wrongly written characters can be improved, and the accuracy of language expression logic for analyzing the text data to be analyzed is improved.
Thirdly, the service problem attribution model determines a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed, the pinyin information of the text data to be analyzed and a service knowledge graph associated with the text data to be analyzed.
Will be divided intoAnalyzing text data, inputting the text data into a service problem attribution model, and obtaining text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise text embedding vectors corresponding to each text unit in the text data to be analyzed, pinyin embedding vectors corresponding to each text unit, paragraph embedding vectors of each text unit in the text data to be analyzed, and embedding vectors of each text unit at the position in the text data to be analyzed. For example, in FIG. 1, E Rice Is a word-embedded vector corresponding to the word of meal, E A Is a paragraph embedding vector of the word meal in the text data to be analyzed of 'cockroach in the meal'. E 3 Is the position embedding vector of the word of rice in the text data to be analyzed of' cockroach in the rice fan Is a pinyin embedded vector corresponding to the meal.
Determining a service knowledge graph embedding vector associated with the text data to be analyzed according to the text data embedding vector to be analyzed, wherein the service knowledge graph embedding vector comprises: the method comprises the steps that a text embedding vector corresponding to each text unit in a service knowledge graph, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph.
When the service problem attribution model analyzes the text data to be analyzed, the frequency of wrongly written characters is reduced by combining the pinyin embedded vector on the basis of determining the text embedded vector, the paragraph embedded vector and the position embedded vector of the text data to be analyzed. And simultaneously, analyzing the language expression logic of the text data to be analyzed by combining the service knowledge graph associated with the text data to be analyzed. The pinyin embedded vector and the service knowledge map are added, so that the accuracy of analyzing the language expression logic of the text data to be analyzed can be improved, and the service problem attribution model can extract the feature data used for analyzing the service problem classification result in the text data to be analyzed according to the language expression logic of the text data to be analyzed. According to the characteristic data, firstly, a service problem big classification result corresponding to the characteristic data is determined in a service problem big classification result list of the service problem attribution model, and then a service problem small classification result corresponding to the characteristic information is selected from a service problem small classification list of the service problem big classification result based on the determined service problem big classification result. And determining a service problem classification result corresponding to the text data to be analyzed according to the determined target service problem large classification result and the target service problem small classification result in the target service problem large classification result.
In addition, the service problem attribution model is obtained by training the following method:
first, a pre-training model for analyzing feature information of text data is obtained.
The pre-training model may be obtained by training a conventional Bert (Bidirectional Encoder with reconstruction from transforms) model. The Bert model is a pre-trained language model and can be used in tasks such as question-answering systems, emotion analysis, spam filtering, named entity recognition, document clustering and the like.
The basic concept of the Bert model is introduced below:
the input object of the Bert model is an original vector of each word in the text to be recognized, and the method comprises the following steps: the method comprises the steps of word vector (Token Embedding), paragraph vector (Segment Embedding) and Position vector (Position Embedding), and inputting a Bert model after directly adding the word vector, the paragraph vector and the Position vector. The paragraph vector is used for separating the context of the text information to be recognized by using the embedded information, and the position vector is the position information of the word in the text to be recognized.
The Bert model includes at least three types of inputs and corresponds to three types of outputs. The method comprises the following specific steps:
1) single text classification: such as emotional analysis of articles. The Bert model adds a [ CLS ] symbol vector in the input paragraph vector, and takes the output vector corresponding to the symbol as the semantic representation of the final whole article. Compared with other characters, the [ CLS ] symbolic vector can more fairly fuse language expression logics of words.
2) Sentence-to-text classification: such as a question and answer system. The input vector of the method is added with [ CLS ] symbolic vector and [ SEP ] symbolic vector for dividing sentences.
3) Sequence labeling classification: the output vector corresponding to each word of such a task is a label for the word, which can be understood as a classification.
In the embodiment of the application, the Transformer structure in the Bert model is obtained through bidirectional information of a multi-head attention mechanism and a covering mechanism, so that the language expression logic of the text data to be recognized is enhanced. The attention mechanism is to let the neural network focus on a portion of the input. The multi-head attention mechanism is to linearly combine a plurality of enhanced semantic vectors of each word to finally obtain an enhanced semantic vector with the length equal to that of the original word vector. For example, "do this package/good" and "do this package/good", whether "eat" differs from the semantic meaning expressed by the "good" combination.
Traditional language models are left to right entering a text sequence, or combining left-to-right and right-to-left training. The two-way information acquisition of the Bert model has an advantage in understanding the context over the one-way language model.
The masking mechanism is that when the text to be recognized is input into the model, any character or any word in a sentence is randomly covered or replaced, and the model predicts the content of the covered or replaced part of the sentence through the context understanding. And then, the model determines attribution classification models corresponding to the text to be recognized according to the emotion reflected by the context content of the text to be recognized after the recognized covered or replaced content is recognized.
In the application, compared with the traditional Bert model, when the pre-training model converts text data to be analyzed into embedded vectors, pinyin embedded vectors are added on the basis of original text embedded vectors, paragraph embedded vectors and position embedded vectors so as to improve the problem of error rate of character recognition. And then embedding a service knowledge graph in the embedded vector of the text data to be analyzed so as to improve the accuracy of analyzing the language expression logic of the text data to be analyzed. Therefore, the pre-training model is used for acquiring the language expression logic of the text data to be analyzed, analyzing and determining the characteristic information of the text data to be analyzed, and making a basis for subsequently determining the service problem classification result of the text data to be analyzed.
And secondly, adjusting the pre-training model according to a text data sample aiming at service information and a service problem big classification result sample aiming at the text data sample to obtain a big classification text data characteristic analysis model, wherein the big classification text data characteristic analysis model is used for analyzing the characteristic information of the text data corresponding to the service problem big classification result.
The training purpose of the pre-training model is to obtain the characteristic information of the text data to be analyzed, and in order to train the pre-training model on the service problem large-classification result parameters, the pre-training model is subjected to parameter training by adopting a text data sample aiming at the service information and a service problem large-classification result sample aiming at the text data sample.
Training the obtained large-classification text data characteristic analysis model, and obtaining a large-classification result of the service problem corresponding to the text data to be analyzed according to the text data to be analyzed. Fig. 2 is a schematic diagram of a service problem classification result for text information in a service problem attribution model provided in the embodiment of the present application. The major classification results of service problems can be foreign bodies, deterioration, food-borne diseases, overdue, uncooked food, and the like.
And thirdly, adjusting the large-classification text data characteristic analysis model according to the text data sample aiming at the service information and the small-classification result sample aiming at the service problem of the text data sample to obtain a service problem attribution model for analyzing the classification result aiming at the service problem of the text data to be analyzed.
The large-classification text data feature analysis model is used for analyzing a large-classification result of the service problems of the text data, and on the basis, the large-classification text data feature analysis model is further subjected to parameter adjustment so as to analyze a small-classification result of the service problems of the text data.
In the second step, the pre-training model is adjusted by adopting a text data sample aiming at service information and a service problem large-classification result sample aiming at the text data sample, so as to obtain a large-classification text data characteristic analysis model. The term "service problem big classification result sample for the text data sample" refers to a service problem big classification result sample belonging to the text data sample, which is used as a positive sample.
In addition, the pre-training model can be trained according to a text data sample aiming at service information and a service problem large-classification result sample which does not belong to the text data sample as a first negative sample pair.
For example, the text data sample is "a cockroach is in this meal", and the large classification result sample of the service problem belonging to the text data sample is "foreign matter", which is a positive sample. The service issue large classification result sample that does not belong to the text data sample may be "spoiled", or "food-borne disease", or "expired", or "uncooked food", etc., which is a negative sample.
The positive sample is a service problem large-class result sample belonging to the text data sample and capable of being inquired by the large-class text data feature analysis model needing training, and the negative sample is a comparison sample of the positive sample, so that the large-class text data feature analysis model can identify a service problem large-class result not belonging to the text data sample. The big classification text data feature analysis model is trained by adopting the positive sample and the negative sample, so that the classification accuracy of the big classification text data sample data feature analysis model on the big classification result of the service problem corresponding to the text data is improved.
In the third step, a text data sample aiming at service information and a small classification result sample aiming at the service problem of the text data sample are adopted, and the large classification text data characteristic analysis model is adjusted to obtain a service problem attribution model. Here, the "sample of the service problem small classification result for the text data sample" mentioned herein refers to the result of the service problem small classification belonging to the text data sample, which is taken as a positive sample. And, here, the service issue small classification result belonging to the text data sample is at least one of a plurality of candidate service issue small classification results among the service issue large classification results determined in the second step.
In addition, the large-classification text data feature analysis model can be trained according to a text data sample aiming at service information and a small-classification result sample of the service problem which does not belong to the text data sample as a second negative sample pair.
For example, the text data sample is "a cockroach exists in the meal", the large classification result sample of the service problem belonging to the text data sample is "a foreign object", and the small classification result sample of the service problem belonging to the text data sample is "a cockroach". The service problem small classification result sample that does not belong to the text data sample may be "sharp foreign matter", or "mouse", or the like, which is a negative sample.
And training the large-classification text data feature analysis model by adopting a positive sample and a negative sample, so that the classification accuracy of the obtained service problem attribution model on the service problem classification result corresponding to the text data is improved.
The embodiment of the application provides a service problem attribution method, which comprises the following steps: acquiring text data to be analyzed aiming at service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
In one embodiment of the application, the service problem attribution model analyzes and obtains a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed. The service problem attribution model provided by the method takes the text data to be analyzed and the service knowledge graph related to the text data to be analyzed as input information, and the service knowledge graph can acquire the attribute information of the service information in the text data to be analyzed, so that the comprehension degree of the language expression logic of the text data to be analyzed is improved. The service knowledge graph is beneficial to improving the accuracy of the service problem attribution model for determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model for matching the text data to be analyzed with the service problem classification result is improved.
First embodiment
A specific flowchart of a service problem attribution method provided in the first embodiment of the present application is shown in fig. 8, which is a flowchart of a service problem attribution method provided in the first embodiment of the present application. The service problem attribution method shown in FIG. 8 comprises the following steps: step S801 to step S802.
As shown in fig. 8, in step S801, text data to be analyzed for service information is obtained.
This step is used to obtain text data to be analyzed for the service information. And after the server side obtains the text data to be analyzed, the server side is a data basis for determining the service problem classification result corresponding to the text data to be analyzed by the service problem attribution model in the subsequent step.
The service information can be food service information for food service provided by a merchant to a user; the analysis text data for the service information can be evaluation information of a user for the food service information; the obtaining of the text data to be analyzed for the service information includes: and acquiring text data to be analyzed aiming at the food service information sent by the user side.
For example, the service information takes a meal provided by a merchant for a user as an example, and the text data to be analyzed of the service information is evaluation information of the meal for the user.
As shown in fig. 8, in step S802, the text data to be analyzed is input into a service problem attribution model, and a service problem classification result for the text data to be analyzed is obtained.
The step is used for obtaining a service problem classification result corresponding to the text data to be analyzed according to the service problem attribution model. A first method for obtaining a service problem classification result for text data to be analyzed by the service problem attribution model according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed is described below.
The service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed. The service knowledge spectrogram is helpful for understanding the language expression logic of the text data to be analyzed, and the accuracy of analyzing the language expression logic of the text data to be analyzed is improved. Therefore, the text data to be analyzed is input into the service problem attribution model, and the service problem attribution model determines the service problem classification result corresponding to the text data to be analyzed according to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed.
It should be noted that, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes:
step 1-1: inputting the text data to be analyzed into a service problem attribution model, and obtaining a first embedded vector aiming at the text data to be analyzed, wherein the first embedded vector comprises a text data to be analyzed embedded vector and a service knowledge graph embedded vector related to the text data to be analyzed;
step 1-2: and acquiring a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector.
Wherein, in step 1-1, the inputting the text data to be analyzed into a service problem attribution model to obtain a first embedded vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed; and obtaining a first embedded vector aiming at the text data to be analyzed according to the embedded vector of the text data to be analyzed and the embedded vector of the service knowledge graph.
It should be noted that the service problem attribution model obtains text data to be analyzed, obtains a service knowledge graph related to the text data to be analyzed, and analyzes and determines a service problem classification result of the text data to be analyzed by using the text data to be analyzed and the service knowledge graph as input information. The service problem attribution model converts the data form of the text data to be analyzed into the text data embedding vector to be analyzed, and converts the data form of the service knowledge graph into the service knowledge graph embedding vector. And generating a first embedded vector aiming at the text data to be analyzed according to the text data embedded vector to be analyzed and the service knowledge map embedded vector related to the text data to be analyzed.
The service knowledge graph related to the text data to be analyzed can be obtained by inquiring the service knowledge graph related to the text data to be analyzed from a service problem attribution model pre-stored service knowledge graph library, and can also be obtained from other sharing platforms.
Inputting the text data to be analyzed into the service problem attribution model to obtain text data embedding vectors to be analyzed, wherein the text data to be analyzed comprises: inputting the text data to be analyzed into the service problem cause model, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, wherein each text unit embeds a vector in a corresponding paragraph in the text data to be analyzed, and each text unit embeds a vector in a corresponding position in the text data to be analyzed; embedding a text embedding vector corresponding to each text unit in the text data to be analyzed, and embedding a vector corresponding to each text unit in the text data to be analyzed at a corresponding position in the text data to be analyzed, and processing to obtain the text data embedding vector to be analyzed.
Wherein, the querying a service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed comprises: acquiring a service information embedded vector in the text data to be analyzed according to the text data embedded vector to be analyzed; according to the service information embedding vector in the text data to be analyzed, inquiring a service knowledge graph aiming at the service information; acquiring a text embedding vector corresponding to each text unit in a service knowledge graph aiming at the service information, wherein each text unit is embedded with a vector in a corresponding paragraph in the service knowledge graph, and each text unit is embedded with a vector in a corresponding position in the service knowledge graph; embedding a text corresponding to each text unit in a service knowledge graph aiming at the service information into a vector, embedding each text unit into a corresponding paragraph in the service knowledge graph, and embedding each text unit into a vector at a corresponding position in the service knowledge graph for processing to obtain the service knowledge graph embedding vector aiming at the service information.
In step 1-2, the service problem classification result corresponding to the text data to be analyzed is obtained according to the first embedded vector, and may be obtained at least in the following three ways, which are discussed below respectively.
The service problem classification result corresponding to the text data to be analyzed is obtained according to the first embedded vector, and may be obtained according to a first method as follows:
acquiring characteristic information of the text data to be analyzed according to the first embedded vector; and according to the characteristic information of the text data to be analyzed, inquiring a target service problem classification result matched with the characteristic information of the text data to be analyzed in a service problem classification list of the service problem attribution model to serve as a service problem classification result corresponding to the text data to be analyzed.
The querying, according to the feature information of the text data to be analyzed, a target service problem classification result matched with the feature information of the text data to be analyzed in a service problem classification list of the service problem attribution model, as a service problem classification result corresponding to the text data to be analyzed, includes: acquiring characteristic information corresponding to a plurality of candidate service problem classification results in a service problem classification list of the service problem attribution model; and comparing the feature information of the text data to be analyzed with the feature information respectively corresponding to the candidate service problem classification results, and determining the candidate service problem classification result containing the feature information of the text data to be analyzed as a target service problem classification result matched with the feature information of the text data to be analyzed as the service problem classification result corresponding to the text data to be analyzed.
For example, the text data to be analyzed is "mildew appears on fruit in this fruit drink", and the characteristic information of the text data to be analyzed is determined as "fruit and mildew" based on the first embedded vector.
The first embedding vector comprises a text data embedding vector to be analyzed and a service knowledge map embedding vector, wherein the text data embedding vector to be analyzed comprises a text unit embedding vector, a position embedding vector and a paragraph embedding vector. The first embedding vector is an embedding vector obtained by combining the text embedding vector to be analyzed and the service knowledge map embedding vector. Therefore, according to the feature information of the text data to be analyzed of the first embedded vector, the comprehensive semantics of the text data to be analyzed needs to be determined by combining the semantics of the text data to be analyzed and the service knowledge map corresponding to the text data to be analyzed.
Here, the text data to be analyzed "the mildew appears on the fruit in the fruit beverage" the corresponding service knowledge map is "the fruit belongs to the fresh fruit and vegetable, and the mildew appears on the surface of the fresh fruit and vegetable" the mildew phenomenon. Therefore, from the first embedded vector, the characteristic information of the text data to be analyzed is "fruit and mildew".
And a plurality of candidate service problem classification results in the service problem classification list, such as two candidate service problem classification results of ' visible mildew and rot ' and ' shown in fig. 2. Wherein, the characteristic information of 'visible mildewing and deterioration' comprises 'the characteristic of mildewing on the surface of the object', and the characteristic information of 'rotting' comprises 'a large number of lesions in the object'. And comparing the characteristic information of the text data to be analyzed with the characteristic information of the two candidate service problem classification results to obtain the text data to be analyzed, wherein the service problem classification result corresponding to the fact that mildew appears on the fruit in the fruit beverage is visible mildewing and deterioration.
(II) obtaining a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector, wherein the service problem classification result can be obtained according to a second mode as follows:
acquiring characteristic information of the text data to be analyzed according to the first embedded vector; according to the characteristic information of the text data to be analyzed, at least one service problem classification result matched with the characteristic information of the text data to be analyzed is inquired in a service problem classification list of the service problem attribution model; and taking the at least one service problem classification result as a service problem classification result corresponding to the text data to be analyzed.
(III) the service problem classification result corresponding to the text data to be analyzed is obtained according to the first embedded vector, and can be obtained according to the following third mode:
acquiring characteristic information of the text data to be analyzed according to the first embedded vector; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
The service problem attribution model comprises a large-class text data characteristic analysis model, and the large-class text data characteristic analysis model is used for analyzing characteristic information of text data corresponding to a large-class result of a service problem; the querying, according to the feature information of the text data to be analyzed, a service problem large classification result associated with the feature information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model, as a target service problem large classification result of the text data to be analyzed, includes: according to the large classification text data feature analysis model, first feature information of text data used for analyzing the large classification result of the service problem in the text data to be analyzed is obtained; acquiring first candidate characteristic information of the text data corresponding to each candidate service problem large classification result in the service problem large classification list; and comparing the first characteristic information in the text data to be analyzed with the first candidate characteristic information corresponding to each candidate service problem large classification result to obtain a target service problem large classification result corresponding to the text data to be analyzed.
On the basis of determining the target service problem large classification result of the text data to be analyzed, querying a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result, including: acquiring second characteristic information of the text data used for analyzing the small classification result of the service problem in the text data to be analyzed; acquiring second candidate characteristic information of the text data corresponding to each candidate service problem small classification result in the service problem small classification list; and comparing the second characteristic information in the text data to be analyzed with the second candidate characteristic information corresponding to each candidate service problem small classification result to obtain a target service problem small classification result corresponding to the text data to be analyzed.
The first method is that the service problem attribution model obtains a service problem classification result for the text data to be analyzed according to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed. According to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed, the accuracy of the language expression logic for analyzing the text data to be analyzed is improved, and on the basis, the accuracy of the service problem classification result corresponding to the text data to be analyzed is improved.
The first method is to improve the accuracy of the linguistic expression logic for analyzing the text data to be analyzed based on the service knowledge graph. In addition, the service problem attribution model further comprises a second mode for obtaining the service problem classification result corresponding to the text data to be analyzed:
the service problem attribution model is specifically used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed, the pinyin information corresponding to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
In the second mode, the service problem classification result corresponding to the text data to be analyzed is analyzed, the service knowledge map of the text data to be analyzed is combined, the pinyin information of the text data to be analyzed is also combined, the problems of wrongly written characters and the like of the text data to be analyzed are solved, and the accuracy of analyzing the language expression logic of the text data to be analyzed is also improved.
Inputting the text data to be analyzed into a service problem attribution model, and obtaining a service problem classification result for the text data to be analyzed, wherein the service problem classification result comprises:
step 2-1: inputting the text data to be analyzed into a service problem attribution model, and obtaining a second embedded vector aiming at the text data to be analyzed, wherein the second embedded vector comprises a pinyin embedded vector of the text data to be analyzed and a pinyin embedded vector of a service knowledge graph related to the text data to be analyzed;
step 2-2: and acquiring a service problem classification result corresponding to the text data to be analyzed according to the second embedded vector.
Wherein, in step 2-1, the inputting the text data to be analyzed into the service problem attribution model and obtaining a second embedded vector for the text data to be analyzed comprises: inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise pinyin embedding vectors corresponding to each text unit in the text data to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed, wherein the service knowledge map embedding vector comprises pinyin embedding vectors corresponding to each text unit in the service knowledge map; and obtaining a second embedded vector aiming at the text data to be analyzed according to the text data embedded vector to be analyzed and the service knowledge map embedded vector.
It should be noted that the service problem attribution model converts the data form of the text data to be analyzed into the form of the embedded vector, and obtains the pinyin embedded vector of the text unit in addition to the text unit embedded vector, the position embedded vector, and the paragraph embedded vector of the text data to be analyzed. The text data to be analyzed "cockroach in this meal" in fig. 1 includes a word embedding vector, a paragraph embedding vector, a position embedding vector, and a pinyin embedding vector for each word. The addition of the pinyin embedded vector of the text unit can improve the probability of wrongly written characters when identifying the characters of the text data to be analyzed. Then, combining the text data embedding vector to be analyzed and the service knowledge map embedding vector related to the text data to be analyzed to generate a second embedding vector for the text data to be analyzed. In the process, the pinyin embedded vector is combined with the service knowledge map, so that the character recognition accuracy in the text data to be analyzed is improved, and the accuracy of analyzing the language logic expression of the text data to be analyzed is also improved.
The second embedded vector obtained in step 201 is obtained by vector processing the text data to be analyzed embedded vector and the service knowledge map embedded vector, and the text data to be analyzed embedded vector in this step includes the pinyin embedded vector of the text data to be analyzed.
The step of inputting the text data to be analyzed into the service problem attribution model to obtain the text data embedding vector to be analyzed comprises the following steps: inputting the text data to be analyzed into the service problem attribution model, obtaining text embedded vectors corresponding to each text unit in the text data to be analyzed, pinyin embedded vectors corresponding to each text unit, paragraph embedded vectors corresponding to each text unit in the text data to be analyzed, and embedded vectors corresponding to each text unit in the text data to be analyzed; embedding a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a position embedding vector corresponding to each text unit in the text data to be analyzed, and processing to obtain the text data embedding vector to be analyzed.
The querying a service knowledge graph embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed comprises: acquiring a service information embedded vector in the text data to be analyzed according to the text data embedded vector to be analyzed; inquiring a service knowledge graph aiming at the service information according to the service information embedded vector in the text data to be analyzed; acquiring a text embedded vector corresponding to each text unit in a service knowledge graph aiming at the service information, a pinyin embedded vector corresponding to each text unit, a paragraph embedded vector corresponding to each text unit in the service knowledge graph, and a position embedded vector corresponding to each text unit in the service knowledge graph; embedding a text corresponding to each text unit in a service knowledge graph aiming at the service information into a vector, embedding a pinyin corresponding to each text unit into a vector, embedding each text unit into a corresponding paragraph in the service knowledge graph, and embedding each text unit into a vector at a corresponding position in the service knowledge graph for processing to obtain the service knowledge graph embedding vector aiming at the service information.
The service problem classification result of the text data to be analyzed is obtained by the service problem attribution model through a second mode. Specifically, when the service problem attribution model converts the text data to be analyzed into the embedded vector, the problem of wrongly written or mispronounced characters in the text data to be analyzed is reduced by combining the pinyin embedded vector corresponding to each text unit on the basis of originally obtaining the text embedded vector, the paragraph embedded vector and the position embedded vector. Meanwhile, the accuracy of the language expression logic for analyzing the text data to be analyzed is improved by combining the service knowledge graph corresponding to the text data to be analyzed. And obtaining a service problem big classification result corresponding to the text data to be analyzed on the basis of determining the language expression logic of the text data to be analyzed. And continuously matching the small service problem classification results corresponding to the text data to be analyzed in the plurality of candidate small service problem classification lists of the determined large service problem classification results, and generating the service problem classification results corresponding to the text data to be analyzed according to the large service problem classification results and the small service problem classification results.
The embodiment of the application provides a service problem attribution method, which comprises the following steps: acquiring text data to be analyzed aiming at service information; inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
According to the method, the service problem attribution model analyzes and obtains the service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and the service knowledge graph related to the text data to be analyzed. According to the service problem attribution model provided by the method, the text data to be analyzed and the service knowledge graph related to the text data to be analyzed are used as input information, and the service knowledge graph can acquire attribute information of service information in the text data to be analyzed, so that the comprehension degree of linguistic expression logic of the text data to be analyzed is improved. The service knowledge graph is beneficial to improving the accuracy of the service problem attribution model for determining the language expression logic of the text data to be analyzed, and the accuracy of the service problem attribution model for matching the text data to be analyzed with the service problem classification result is improved.
Second embodiment
Corresponding to the embodiments corresponding to the application scenarios of the service problem attribution method provided by the present application and the service problem attribution method provided by the first embodiment, a second embodiment of the present application also provides a service problem attribution device. Since the embodiment of the apparatus is basically similar to the embodiment corresponding to the application scenario and the first embodiment, the description is relatively simple, and for relevant points, reference may be made to the embodiment corresponding to the application scenario and part of the description of the first embodiment. The device embodiments described below are merely illustrative.
Please refer to fig. 9, which is a schematic diagram of a service problem attribution apparatus according to a second embodiment of the present application. A service problem attribution apparatus provided in a second embodiment of the present application, includes:
a first obtaining unit 901, configured to obtain text data to be analyzed for service information.
A second obtaining unit 902, configured to input the text data to be analyzed into a service problem attribution model, and obtain a service problem classification result for the text data to be analyzed; the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
Third embodiment
Corresponding to the embodiments corresponding to the application scenarios of the service problem attribution method provided by the present application and the service problem attribution method provided by the first embodiment, the third embodiment of the present application further provides another service problem attribution method.
Please refer to fig. 10, which is a schematic diagram illustrating another service problem attribution method according to a third embodiment of the present application. The service problem attribution method shown in FIG. 10 comprises the following steps: step S1001 to step S1002.
As shown in fig. 10, in step S1001, text data to be analyzed for service information is obtained.
As shown in fig. 10, in step S1002, the text data to be analyzed is input into a service problem attribution model, and a service problem classification result for the text data to be analyzed is obtained.
The service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise text embedding vectors corresponding to each text unit in the text data to be analyzed, pinyin embedding vectors corresponding to each text unit, paragraph embedding vectors corresponding to each text unit in the text data to be analyzed, and embedding vectors corresponding to each text unit in the position of the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain an embedded vector of the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, carrying out vectorization processing on the text data to be analyzed, and obtaining a text embedding vector corresponding to each text unit in the text data to be analyzed, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector corresponding to each text unit in the text data to be analyzed, and a vector embedded in a corresponding position of each text unit in the text data to be analyzed; embedding a vector into a text corresponding to each text unit in the text data to be analyzed, embedding a vector into a pinyin corresponding to each text unit, embedding a vector into a paragraph corresponding to each text unit in the text data to be analyzed, and embedding a vector into a position corresponding to each text unit in the text data to be analyzed for processing to obtain the text data embedding vector to be analyzed.
Optionally, the obtaining a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the text data embedding vector to be analyzed; and according to the characteristic information of the text data to be analyzed, inquiring a target service problem classification result matched with the characteristic information of the text data to be analyzed in a service problem classification list of the service problem attribution model to serve as a service problem classification result corresponding to the text data to be analyzed.
Optionally, the obtaining a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed includes: acquiring characteristic information of the text data to be analyzed according to the text data embedding vector to be analyzed; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
Optionally, the service problem attribution model is specifically configured to obtain a service problem classification result for the text data to be analyzed according to the text data to be analyzed, pinyin information corresponding to the text data to be analyzed, and a service knowledge graph related to the text data to be analyzed.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed includes: inputting the text data to be analyzed into a service problem attribution model, and obtaining a second embedded vector aiming at the text data to be analyzed, wherein the second embedded vector comprises a text data to be analyzed embedded vector and a service knowledge graph embedded vector related to the text data to be analyzed; and acquiring a service problem classification result corresponding to the text data to be analyzed according to the second embedded vector.
Optionally, the inputting the text data to be analyzed into a service problem attribution model to obtain a second embedded vector for the text data to be analyzed includes: inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise pinyin embedding vectors corresponding to each text unit in the text data to be analyzed; inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed, wherein the service knowledge map embedding vector comprises pinyin embedding vectors corresponding to each text unit in the service knowledge map; and obtaining a second embedded vector aiming at the text data to be analyzed according to the embedded vector of the text data to be analyzed and the embedded vector of the service knowledge graph.
Optionally, the obtaining, according to the second embedded vector, a service problem classification result corresponding to the text data to be analyzed includes: acquiring feature information of the text data to be analyzed according to the second embedded vector; according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed; on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result; and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
Fourth embodiment
Corresponding to the embodiment corresponding to the application scenario of the service problem attribution method provided by the application and the service problem attribution method provided by the third embodiment, the fourth embodiment of the application also provides another service problem attribution device. Since the embodiment of the apparatus is basically similar to the embodiment and the third embodiment corresponding to the application scenario, the description is relatively simple, and for the relevant points, reference may be made to the embodiment and the third embodiment corresponding to the application scenario for partial description. The device embodiments described below are merely illustrative.
Please refer to fig. 11, which is a schematic diagram of another service problem attribution device provided in a fourth embodiment of the present application. Another service problem attribution apparatus provided in the fourth embodiment of the present application, includes:
a third obtaining unit 1101, configured to obtain text data to be analyzed for the service information.
A fourth obtaining unit 1102, configured to input the text data to be analyzed into a service problem attribution model, and obtain a service problem classification result for the text data to be analyzed. The service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.
Fifth embodiment
Corresponding to the embodiments corresponding to the application scenarios of the service problem attribution method provided by the present application and the service problem attribution method provided by the first embodiment, a fifth embodiment of the present application further provides a training method of a service problem attribution model.
Please refer to fig. 12, which is a schematic diagram illustrating a training method of a service problem attribution model according to a fifth embodiment of the present application. The service problem attribution method shown in fig. 12 comprises the following steps: step S1201 to step S1203.
As shown in fig. 12, in step S1201, a pre-trained model for analyzing feature information of text data is obtained.
As shown in fig. 12, in step S1202, the pre-training model is adjusted according to a text data sample for service information and a service problem large classification result sample for the text data sample, so as to obtain a large classification text data feature analysis model, where the large classification text data feature analysis model is used for analyzing feature information of text data corresponding to the service problem large classification result.
As shown in fig. 12, in step S1203, the large-category text data feature analysis model is adjusted according to the text data sample for the service information and the small service problem classification result sample for the text data sample, so as to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.
Optionally, the pre-training model analyzes feature information of the text data in the following manner: obtaining text data embedding vectors according to the text data, wherein the text data embedding vectors comprise text embedding vectors corresponding to each text unit in the text data, pinyin embedding vectors corresponding to each text unit, paragraph embedding vectors of each text unit in the text data, and position embedding vectors of each text unit in the text data; and analyzing the characteristic information of the text data according to the text data embedded vector.
Optionally, the method further includes: acquiring a service knowledge graph embedding vector associated with the text data according to the text data embedding vector, wherein the service knowledge graph embedding vector comprises a text embedding vector corresponding to each text unit in a service knowledge graph associated with the text data, a pinyin embedding vector corresponding to each text unit, a paragraph embedding vector of each text unit in the service knowledge graph, and a position embedding vector of each text unit in the service knowledge graph; analyzing the feature information of the text data according to the text data embedded vector, wherein the analyzing comprises: analyzing feature information of the text data according to the text data embedding vector and a service knowledge graph embedding vector associated with the text data.
Optionally, the adjusting the pre-training model according to the text data sample for the service information and the sample for the large classification result of the service problem for the text data sample to obtain a large classification text data feature analysis model includes: inputting the text data sample aiming at the service information into the pre-training model to obtain a first service problem large classification result of the text data sample aiming at the service information output by the pre-training model; and adjusting the service problem major classification parameters of the pre-training model according to the similarity between the first service problem major classification result and the service problem major classification result sample aiming at the text data sample to obtain the major classification text data feature analysis model.
Optionally, the inputting the text data sample for the service information into the pre-training model to obtain a first service problem large classification result of the text data sample for the service information output by the pre-training model includes: inputting the text data sample aiming at the service information into the pre-training model to obtain the characteristic information of the text data sample; and obtaining a first service problem big classification result of the text data sample aiming at the service information and output by the pre-training model according to the characteristic information of the text data sample.
Optionally, the adjusting the large-category text data feature analysis model according to the text data sample for the service information and the small service problem classification result sample for the text data sample to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed includes: inputting text data aiming at the service information into the large-class text data feature analysis model to obtain a second small service problem classification result aiming at the text data sample and output by the large-class text data feature analysis model; and adjusting the large-classification text data feature analysis model according to the similarity degree between the second small service problem classification result and the small service problem classification result sample to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed.
Optionally, the method further includes: taking a text data sample aiming at service information and a service problem large-classification result sample which does not belong to the text data as a first negative sample pair, and adjusting the pre-training model to obtain a large-classification text data feature analysis model for analyzing feature information of the text data corresponding to the service problem large-classification result; and taking a text data sample aiming at the service information and a service problem small classification result sample which does not belong to the text data as a second negative sample pair, and adjusting the large classification text data feature analysis model to obtain a service problem attribution model for analyzing the service problem classification result aiming at the text data to be analyzed.
Sixth embodiment
Corresponding to the embodiments corresponding to the application scenarios of the service problem attribution method provided by the present application and the training method of the service problem attribution model provided by the fifth embodiment, the sixth embodiment of the present application further provides a training device of the service problem attribution model. Since the device embodiment is basically similar to the embodiment corresponding to the application scenario and the fifth embodiment, the description is relatively simple, and for relevant points, reference may be made to the embodiment corresponding to the application scenario and part of the description of the fifth embodiment. The device embodiments described below are merely illustrative.
Please refer to fig. 13, which is a schematic diagram of a training apparatus for a service problem attribution model according to a sixth embodiment of the present application. A training apparatus for a service problem attribution model provided in a sixth embodiment of the present application includes:
the pre-training model obtaining unit 1301 is configured to obtain a pre-training model for analyzing feature information of text data.
A large-category text data feature analysis model obtaining unit 1302, configured to adjust the pre-training model according to a text data sample for service information and a large-category service problem result sample for the text data sample, so as to obtain a large-category text data feature analysis model, where the large-category text data feature analysis model is used to analyze feature information of text data corresponding to the large-category service problem result.
And a service problem attribution model obtaining unit 1303, configured to adjust the large-class text data feature analysis model according to the text data sample for the service information and the small-class service problem result sample for the text data sample, so as to obtain a service problem attribution model for analyzing the service problem classification result for the text data to be analyzed.
Seventh embodiment
Corresponding to the above method embodiments provided by the present application, a seventh embodiment of the present application further provides an electronic device. Since the seventh embodiment is substantially similar to the method embodiment provided in this application, the description is relatively simple, and reference may be made to the description of the method embodiment provided in this application for relevant points. The seventh embodiment described below is merely illustrative.
Please refer to fig. 14, which is a schematic diagram of an electronic device according to a seventh embodiment of the present application. The electronic device includes: at least one processor 1401, at least one communication interface 1402, at least one memory 1403, and at least one communication bus 1404; optionally, the communication interface 1402 may be an interface of a communication module, such as an interface of a WLAN (Wireless Local Area Network) module; the processor 1401 may be a processor CPU or an application Specific Integrated circuit (asic) or one or more Integrated circuits configured to implement embodiments of the present application. Memory 1403 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk memory. Wherein, the memory 1403 stores programs, and the processor 1401 calls the programs stored in the memory 1403 to execute the above method provided by the embodiment of the present application.
Eighth embodiment
Corresponding to the above method embodiments provided by the present application, the eighth embodiment of the present application further provides a computer storage medium, and since the eighth embodiment is substantially similar to the above method embodiments provided by the present application, the description is relatively simple, and relevant points can be referred to the description of the above method embodiment section provided by the present application. The eighth embodiment described below is merely illustrative. The computer storage medium stores a computer program that, when executed, implements the methods provided in the above-described method embodiments. It should be noted that, for the detailed description of the storage medium provided in the eighth embodiment of the present application, reference may be made to the description of the foregoing method embodiment provided in the present application, and details are not repeated here.
Although the present application has been described with reference to the preferred embodiments, it is not intended to limit the present application, and those skilled in the art can make variations and modifications without departing from the spirit and scope of the present application, therefore, the scope of the present application should be determined by the claims that follow.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory. The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
1. Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer readable Media does not include non-Transitory computer readable Media (transient Media), such as modulated data signals and carrier waves.
2. As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
Claims (10)
1. A service problem attribution method, comprising:
acquiring text data to be analyzed aiming at service information;
inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed;
the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and a service knowledge graph related to the text data to be analyzed.
2. The method of claim 1, wherein the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed comprises:
inputting the text data to be analyzed into a service problem attribution model, and obtaining a first embedded vector aiming at the text data to be analyzed, wherein the first embedded vector comprises a text data to be analyzed embedded vector and a service knowledge graph embedded vector related to the text data to be analyzed;
and acquiring a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector.
3. The method of claim 2, wherein the inputting the text data to be analyzed into a service problem attribution model, obtaining a first embedding vector for the text data to be analyzed, comprises:
inputting the text data to be analyzed into the service problem cause model to obtain text data embedding vectors to be analyzed;
inquiring a service knowledge map embedding vector associated with the text data embedding vector to be analyzed according to the text data embedding vector to be analyzed;
and obtaining a first embedded vector aiming at the text data to be analyzed according to the text data embedded vector to be analyzed and the service knowledge map embedded vector.
4. The method according to claim 2, wherein the obtaining a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector comprises:
acquiring characteristic information of the text data to be analyzed according to the first embedded vector;
and according to the characteristic information of the text data to be analyzed, inquiring a target service problem classification result matched with the characteristic information of the text data to be analyzed in a service problem classification list of the service problem attribution model to serve as a service problem classification result corresponding to the text data to be analyzed.
5. The method according to claim 2, wherein the obtaining a service problem classification result corresponding to the text data to be analyzed according to the first embedded vector comprises:
acquiring characteristic information of the text data to be analyzed according to the first embedded vector;
according to the characteristic information of the text data to be analyzed, inquiring a service problem large classification result associated with the characteristic information of the text data to be analyzed in a service problem large classification list in the service problem attribution classification model to serve as a target service problem large classification result of the text data to be analyzed;
on the basis of determining a target service problem large classification result of the text data to be analyzed, inquiring a target service problem small classification result associated with the feature information of the text data to be analyzed in a service problem small classification list of the target service problem large classification result;
and generating a service problem classification result corresponding to the text data to be analyzed according to the target service problem large classification result and the target service problem small classification result.
6. The method according to claim 1, wherein the service problem attribution model is specifically configured to obtain a service problem classification result for the text data to be analyzed according to the text data to be analyzed, pinyin information corresponding to the text data to be analyzed, and a service knowledge graph related to the text data to be analyzed.
7. The method of claim 6, wherein the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed comprises:
inputting the text data to be analyzed into a service problem attribution model, and obtaining a second embedded vector aiming at the text data to be analyzed, wherein the second embedded vector comprises a pinyin embedded vector of the text data to be analyzed and a pinyin embedded vector of a service knowledge graph related to the text data to be analyzed;
and acquiring a service problem classification result corresponding to the text data to be analyzed according to the second embedded vector.
8. A service problem attribution method, comprising:
acquiring text data to be analyzed aiming at service information;
inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result aiming at the text data to be analyzed;
the service problem attribution model is used for obtaining a service problem classification result aiming at the text data to be analyzed according to the text data to be analyzed and pinyin information corresponding to the text data to be analyzed.
9. The method of claim 8, wherein the inputting the text data to be analyzed into a service problem attribution model to obtain a service problem classification result for the text data to be analyzed comprises:
inputting the text data to be analyzed into a service problem attribution model to obtain text data embedding vectors to be analyzed, wherein the text data embedding vectors to be analyzed comprise text embedding vectors corresponding to each text unit in the text data to be analyzed, pinyin embedding vectors corresponding to each text unit, paragraph embedding vectors corresponding to each text unit in the text data to be analyzed, and embedding vectors corresponding to each text unit in the position of the text data to be analyzed;
and acquiring a service problem classification result corresponding to the text data to be analyzed according to the text data embedding vector to be analyzed.
10. A training method of a service problem attribution model is characterized by comprising the following steps:
obtaining a pre-training model for analyzing feature information of text data;
adjusting the pre-training model according to a text data sample aiming at service information and a service problem big classification result sample aiming at the text data sample to obtain a big classification text data feature analysis model, wherein the big classification text data feature analysis model is used for analyzing feature information of text data corresponding to the service problem big classification result;
and adjusting the large-classification text data feature analysis model according to the text data sample aiming at the service information and the small-classification result sample aiming at the service problem of the text data sample to obtain a service problem attribution model for analyzing the classification result aiming at the service problem of the text data to be analyzed.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210763456.8A CN115062605A (en) | 2022-06-29 | 2022-06-29 | Service problem attribution method and device |
| PCT/CN2023/103688 WO2024002216A1 (en) | 2022-06-29 | 2023-06-29 | Service problem attribution method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210763456.8A CN115062605A (en) | 2022-06-29 | 2022-06-29 | Service problem attribution method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115062605A true CN115062605A (en) | 2022-09-16 |
Family
ID=83204508
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210763456.8A Pending CN115062605A (en) | 2022-06-29 | 2022-06-29 | Service problem attribution method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN115062605A (en) |
| WO (1) | WO2024002216A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024002216A1 (en) * | 2022-06-29 | 2024-01-04 | 拉扎斯网络科技（上海）有限公司 | Service problem attribution method and device |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021107448A1 (en) * | 2019-11-25 | 2021-06-03 | 주식회사 데이터마케팅코리아 | Method and apparatus for providing knowledge graph-based marketing information analysis service to support efficient document classification processing |
| CN111177416A (en) * | 2020-04-13 | 2020-05-19 | 傲林科技有限公司 | Event root cause analysis model construction method, event root cause analysis method and device |
| CN114429284A (en) * | 2021-12-31 | 2022-05-03 | 拉扎斯网络科技（上海）有限公司 | Method for controlling food risk on line, server, terminal and electronic equipment |
| CN115062605A (en) * | 2022-06-29 | 2022-09-16 | 拉扎斯网络科技（上海）有限公司 | Service problem attribution method and device |
-
2022
- 2022-06-29 CN CN202210763456.8A patent/CN115062605A/en active Pending
-
2023
- 2023-06-29 WO PCT/CN2023/103688 patent/WO2024002216A1/en unknown
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024002216A1 (en) * | 2022-06-29 | 2024-01-04 | 拉扎斯网络科技（上海）有限公司 | Service problem attribution method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2024002216A1 (en) | 2024-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110298725B (en) | Commodity combination recommendation method and device, electronic equipment and readable storage medium | |
| CN110114764B (en) | Providing dietary assistance in conversation | |
| CN110325986B (en) | Article processing method, article processing device, server and storage medium | |
| US9767386B2 (en) | Training a classifier algorithm used for automatically generating tags to be applied to images | |
| CN116821308B (en) | Generation method, training method and device of model and storage medium | |
| CN108694647B (en) | Method and device for mining merchant recommendation reason and electronic equipment | |
| US10671619B2 (en) | Information processing system and information processing method | |
| CN110827797B (en) | Voice response event classification processing method and device | |
| CN107833088A (en) | Content providing, device and smart machine | |
| CN114580405A (en) | Method and device for analyzing commodity comment text, electronic equipment and storage medium | |
| KR20190083143A (en) | Sensory evaluation method and apparatus | |
| KR101319413B1 (en) | Summary Information Generating System and Method for Review of Product and Service | |
| CN115660292A (en) | Carbon emission monitoring method and equipment based on catering consumption data processing | |
| CN117788109A (en) | Method for generating commodity label based on large language model and electronic equipment | |
| CN109214956B (en) | Meal pushing method and device | |
| CN117746143A (en) | AIGC-based image description text generation method and device and storage medium | |
| JP5933863B1 (en) | Data analysis system, control method, control program, and recording medium | |
| CN115062605A (en) | Service problem attribution method and device | |
| CN111666513A (en) | Page processing method and device, electronic equipment and readable storage medium | |
| CN116862626B (en) | Multi-mode commodity alignment method | |
| US11232325B2 (en) | Data analysis system, method for controlling data analysis system, and recording medium | |
| CN113435967B (en) | Method and system for automatically determining marketing content of agricultural products | |
| EP3929931A1 (en) | Dish nutritional value estimation | |
| Hakguder et al. | Smart Diet Management through Food Image and Cooking Recipe Analysis | |
| CN117540003B (en) | Text processing method and related device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |