CN113012774B - Automatic medical record coding method and device, electronic equipment and storage medium - Google Patents
Automatic medical record coding method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113012774B CN113012774B CN201911310848.3A CN201911310848A CN113012774B CN 113012774 B CN113012774 B CN 113012774B CN 201911310848 A CN201911310848 A CN 201911310848A CN 113012774 B CN113012774 B CN 113012774B
- Authority
- CN
- China
- Prior art keywords
- medical data
- historical
- medical
- data
- representation
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 238000013145 classification model Methods 0.000 claims abstract description 35
- 239000013598 vector Substances 0.000 claims description 85
- 238000010801 machine learning Methods 0.000 claims description 62
- 238000012549 training Methods 0.000 claims description 41
- 238000012545 processing Methods 0.000 claims description 14
- 230000011218 segmentation Effects 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 14
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 201000010099 disease Diseases 0.000 description 7
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 7
- 238000003745 diagnosis Methods 0.000 description 5
- 238000013527 convolutional neural network Methods 0.000 description 4
- 208000028659 discharge Diseases 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000003058 natural language processing Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 2
- 239000013307 optical fiber Substances 0.000 description 2
- 230000002980 postoperative effect Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- 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
- G06F16/353—Clustering; Classification into predefined classes
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The embodiment of the disclosure provides an automatic medical record coding method, an automatic medical record coding device, electronic equipment and a computer readable storage medium, which belong to the technical field of medical data, and the automatic medical record coding method comprises the following steps: acquiring medical data to be processed; performing medical record coding classification recognition on the medical data to be processed by combining a medical classification model and a medical attention model to obtain a classification result of the medical data to be processed; and determining the coding information corresponding to the medical data to be processed according to the classification result. The embodiment of the disclosure can automatically determine the coding information of the medical data to be processed, and improve the accuracy of the coding information determination.
Description
Technical Field
The embodiment of the disclosure relates to the technical field of medical data, in particular to an automatic medical record encoding method, an automatic medical record encoding device, electronic equipment and a computer readable storage medium.
Background
The hospital medical records management department can arrange and file each patient medical record, and the international disease classification code is one of the important processes for the patient allocation.
In the related art, coding is generally completed manually according to discharge diagnosis information and other information of electronic medical records according to expert knowledge and referring to a code table. In the mode, the working efficiency is low, the manual knowledge range is required to be more, and the mode has certain dependence and limitation. In addition, when the same data is processed, the obtained results may be different, and thus the consistency and reliability are poor. And may lead to misrecognition, resulting in coding errors and lower accuracy.
It should be noted that the information disclosed in the above background section is only for enhancing understanding of the background of the invention and thus may include information that does not form the prior art that is already known to those of ordinary skill in the art.
Disclosure of Invention
An object of the embodiments of the present disclosure is to provide a medical records automatic coding method, a medical records automatic coding device, an electronic device and a computer readable storage medium, so as to overcome the problems of low efficiency and poor reliability at least to some extent.
Other features and advantages of embodiments of the present disclosure will be apparent from the following detailed description, or may be learned in part by the practice of the invention.
According to an aspect of the embodiments of the present disclosure, there is provided an automatic medical record encoding method, including: acquiring medical data to be processed; performing medical record coding classification recognition on the medical data to be processed by combining a medical classification model and a medical attention model to obtain a classification result of the medical data to be processed; and determining the coding information corresponding to the medical data to be processed according to the classification result.
In an exemplary embodiment of the present disclosure, performing medical record encoding classification recognition on the medical data to be processed in combination with a medical classification model and a medical attention model to obtain a classification result of the medical data to be processed includes: extracting features of the medical data to be processed through the medical classification model to determine feature representation of the medical data to be processed; weighting the characteristic representation based on the medical attention model to obtain a text representation of the medical data to be processed; and determining the classification result according to the text representation of the medical data to be processed.
In an exemplary embodiment of the present disclosure, the method further comprises: training a machine learning model through historical medical data and real coding information of the historical medical data, and taking the trained machine learning model as the medical classification model.
In one exemplary embodiment of the present disclosure, training a machine learning model with historical medical data and encoded information of the historical medical data includes: acquiring word embedding vector representations corresponding to the historical medical data; and training the machine learning model according to the word embedded vector representation and the real coding information of the historical medical data to obtain the trained machine learning model.
In one exemplary embodiment of the present disclosure, obtaining a word embedding vector representation corresponding to the historical medical data includes: acquiring first type data and second type data as the historical medical data, and performing word segmentation on the historical medical data to obtain segmented historical medical data; training the word embedding vector representation according to the segmented historical medical data.
In one exemplary embodiment of the present disclosure, training the machine learning model based on the word embedded vector representation and the true encoded information of the historical medical data to obtain the trained machine learning model includes: determining predictive coding information for the historical medical data based on the word embedded vector representation; and training the machine learning model based on the predictive coding information and the real coding information to obtain the trained machine learning model.
In one exemplary embodiment of the present disclosure, determining predictive coding information for the historical medical data from the word embedding vector representation includes: inputting the word embedded vector representation into the machine learning model to obtain a historical feature representation; configuring coding vectors for a plurality of pieces of coding information, and matching historical characteristic representations corresponding to the historical medical data with the coding vectors to determine matching degree; weighting the historical characteristic representation according to the matching degree to obtain a text representation; the predictive coding information for the historical medical data is determined from the text representation.
In one exemplary embodiment of the present disclosure, matching the historical feature representation corresponding to the historical medical data with the encoding vector to determine a degree of matching includes: and performing first logic operation on the historical characteristic representation and the coding vector to obtain the matching degree between the historical characteristic representation and the coding information.
In an exemplary embodiment of the present disclosure, weighting the historical feature representation according to the degree of matching includes: and performing a second logic operation on the matching degree and each historical characteristic representation to determine the text representation.
According to one aspect of the present disclosure, there is provided an automatic medical records encoding apparatus, including: the data acquisition module is used for acquiring medical data to be processed; the coding classification module is used for carrying out medical record coding classification identification on the medical data to be processed by combining the medical classification model and the medical attention model so as to obtain a classification result of the medical data to be processed; and the code determining module is used for determining the code information corresponding to the medical data to be processed according to the classification result.
According to one aspect of the present disclosure, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the medical records automatic encoding method as set forth in any one of the above.
According to one aspect of the present disclosure, there is provided an electronic device including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the medical records auto-coding method of any one of the above via execution of the executable instructions.
In the medical records automatic coding method, the device, the electronic equipment and the computer readable storage medium provided by the embodiment of the disclosure, the characteristic representation of the medical data to be processed is determined through the medical classification model, and the classification result of the medical data to be processed is further determined through the medical attention model on the basis of the characteristic representation, so that the coding information of the medical data to be processed is determined according to the classification result. On the one hand, the medical data to be processed can be subjected to medical record prediction classification through the model, so that the medical record encoding process of the medical data to be processed is automatically completed, the dependence on the knowledge range and the limitation caused by manual operation are avoided, and the encoding efficiency is improved. On the other hand, the coding information of the medical data to be processed is determined by the same coding rule, so that the problem that the result of the same data is different from person to person in a manual mode is avoided, and the consistency and the reliability are improved. On the other hand, as the medical classification model and the medical attention model are obtained through the training of the historical medical data, and the characteristics of the data are accurately determined through the models and the automatic classification and recognition are carried out, the situation of false recognition can be avoided, and meanwhile, the accuracy of determining the coding information is improved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention. It is evident that the drawings in the following description are only some embodiments of the present invention and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art. In the drawings:
FIG. 1 schematically illustrates a flow diagram of a method for automatic encoding of medical records in accordance with an embodiment of the present disclosure;
FIG. 2 schematically illustrates a flow diagram of determining a trained machine learning model according to an embodiment of the disclosure;
FIG. 3 schematically illustrates a flow diagram of model training of an embodiment of the present disclosure;
FIG. 4 schematically illustrates a flow diagram of determining predictive coding information in an embodiment of the disclosure;
FIG. 5 schematically illustrates a block diagram of a medical records automatic encoding device according to an embodiment of the present disclosure;
fig. 6 schematically shows a block diagram of an electronic device for implementing the above-described automatic medical records encoding method.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. However, the exemplary embodiments may be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art.
Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the invention. One skilled in the relevant art will recognize, however, that the invention may be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other instances, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the invention.
The block diagrams depicted in the figures are merely functional entities and do not necessarily correspond to physically separate entities. That is, the functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
The flow diagrams depicted in the figures are exemplary only, and do not necessarily include all of the elements and operations/steps, nor must they be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the order of actual execution may be changed according to actual situations.
In order to solve the above-mentioned problems, in the embodiments of the present disclosure, an automatic medical record encoding method is first provided, and the automatic medical record encoding method may be applied to an application scenario for processing medical data. The main execution body of the automatic medical record encoding method may be a server, and referring to fig. 1, the automatic medical record encoding method may include step S110, step S120, and step S130. Wherein:
in step S110, medical data to be processed is acquired;
In step S120, performing medical record coding classification recognition on the medical data to be processed by combining a medical classification model and a medical attention model, so as to obtain a classification result of the medical data to be processed;
In step S130, the coding information corresponding to the medical data to be processed is determined according to the classification result.
In the technical scheme provided by the example embodiment of the disclosure, on one hand, the medical data to be processed can be subjected to coding classification identification through the model so as to automatically complete the automatic coding process of the medical data to be processed, so that the dependence on the knowledge range and the limitation caused by the manual operation are avoided, and the coding efficiency is improved. On the other hand, the coding information of the medical data to be processed is determined by the same coding rule, so that the problem that the result of the same data is different from person to person in a manual mode is avoided, and the consistency and the reliability are improved. On the other hand, as the medical classification model and the medical attention model are obtained through the historical medical data training, the situation of false recognition can be avoided, and meanwhile, the accuracy of determining the coding information is improved.
Next, the automatic encoding method of the medical records in the embodiments of the present disclosure will be further explained with reference to the accompanying drawings.
In step S110, medical data to be processed is acquired.
In the embodiment of the disclosure, the medical data to be processed may be data corresponding to the target object. The target object may be a patient in a preset location, which may be, for example, a hospital or a medical center, etc., that can be accessed by the patient. The medical data to be processed may be all medical data associated with the patient, such as all medical record data for each patient. The medical record data is used for recording all records of the target object in the preset place. The medical record data may include a plurality of types of text data, and the plurality of types may include, but are not limited to, any one or more of a complaint and five history of the target object, a diagnosis and treatment process, a diagnosis basis, a post-operative course record, a consultation record, and an examination conclusion.
The medical record data refers to data in medical records generated in the process of seeking medical attention by a target object, and specifically may include clinical medical record data stored in an electronic medical record database. The electronic medical record database may be a data warehouse provided in the terminal device for storing medical data, or may be provided in a server for storing medical data. The server may acquire medical record data from a data warehouse of the terminal device, or may acquire medical record data from a server storing the data. The medical data to be processed may be in the form of text data, such as text entered by a physician in a medical record or manually written, without limitation.
In step S120, a medical record encoding classification is performed on the medical data to be processed in combination with the medical classification model and the medical attention model, so as to obtain a classification result of the medical data to be processed.
In the embodiment of the present disclosure, the medical classification model refers to a model for analyzing medical data to be processed for identifying and predicting which kind of encoded information the medical data to be processed of the target object is. The medical classification model may be any suitable machine learning model capable of classifying, for example, any of decision tree, neural network model, linear regression analysis model, support vector machine, random forest model, and the machine learning model is exemplified herein as a convolutional neural network model. Specifically, the medical classification model is used for performing feature extraction on medical data to be processed to obtain feature data representing main features of the medical data to be processed. The feature data may be described in particular by a feature representation, which refers to a form for representing the feature data. Thus, the features are expressed as outputs obtained through a convolutional neural network.
Medical attention model refers to a network based on an attention mechanism that allows a neural network to focus on only a portion of its inputs and to select a particular input. The attention mechanism may be applied to any type of input, regardless of its shape, for matrix form of input, such as images or vectors, etc. After deriving the feature representation based on the medical classification model, the feature representation derived from the medical classification model may be further processed based on the medical attention model to obtain a textual representation of the medical data to be processed. The text representation refers to feature data that is directly used to determine the coding classification of the medical data to be processed, and the text representation may be associated with the feature representation. In text classification, text needs to be converted into a form that can be processed by a computer algorithm, so the accuracy of text representation directly affects the accuracy of the results of natural language processing. By determining an accurate text representation, the characteristics of the medical data to be processed can be accurately determined, and the accuracy of the classification result can be improved.
In the embodiment of the disclosure, the medical attention model is used for performing weighted change on the medical data to be processed, for example, calculating attention weights of all data in the medical data to be processed, or screening out a part of attention which does not meet the condition after generating the attention weights, making the attention weights of the data to be processed be 0, and so on. Medical attention models are mainly used to focus attention on important data. The principle of the medical attention model is to calculate the matching degree of the current input sequence and the output vector, and the matching degree is high, namely the relative score of the attention concentrating point is higher. The input of the medical attention model may be a text representation corresponding to the feature representation. The medical data to be processed is screened through the medical attention model, and important data can be screened out from the medical data, so that attention is focused on the important data, and the accuracy of data processing is improved. And words or sentences related to the coding information can be obtained from the medical record data, coding basis is given, and the interpretability is increased.
Before using the medical classification model, the machine learning model may be trained first in order to improve its accuracy. The historical medical data with the determined coding information can be used for training, and specifically, the historical medical data and the real coding information of the historical medical data can be used for training a machine learning model until a trained machine learning model is obtained and used as a medical classification model. The historical medical data refers to data of a plurality of objects in the preset place, and the data of the plurality of objects can be historical medical record data of other objects except the target object or can be historical medical record data of the target object.
A flow chart for determining a trained machine learning model is schematically shown in fig. 2, and with reference to fig. 2, mainly comprises the following steps:
In step S210, a word embedding vector representation corresponding to the historical medical data is acquired.
In this step, the word embedding vector refers to a real vector in which words in a vocabulary are mapped to a low dimensional space with respect to the size of the vocabulary, and may be a word vector expressed in a word embedding form. That is, the historical medical data is described in terms of word-embedded vector representations. The word embedded vector representation may be used directly to input the machine learning model. Specifically, step S210 for acquiring the word embedding vector representation may include step S211 and step S212, wherein:
in step S211, acquiring first type data and second type data as the historical medical data, and performing word segmentation on the historical medical data to obtain segmented historical medical data;
in step S212, the word embedding vector representation is trained from the segmented historical medical data.
In the embodiment of the disclosure, the first type of data and the second type of data have different sources, and the first type of data may be data contained in medical record data, such as natural text data including a first complaint and a fifth history, a diagnosis and treatment process, a diagnosis basis, a postoperative course record, a consultation record, an examination conclusion and the like; the second type of data refers to data contained in the discharge record, which may be, for example, diagnostic field data in the discharge record. Based on this, the historical medical data of the plurality of subjects can be obtained from the first type data and the second type data.
Further, the history medical data may be segmented to obtain segmented history medical data. The word segmentation algorithm may be specifically used to segment the historical medical data, and the word segmentation algorithm may include any one of a minimum matching algorithm, a forward (reverse) maximum matching method, a word-by-word matching algorithm and a neural network algorithm, which is not limited herein. By word segmentation of the historical medical data, automatic sentence meaning recognition can be better assisted, and classification accuracy is improved.
When the segmented historical medical data is obtained, training can be carried out on the segmented historical medical data to obtain word embedding vector representation. Specifically, the word2vec model or the glove model may be used to train the segmented historical medical data, and the word2vec model is taken as an example. word2vec is a natural language processing tool that vectorizes all words to quantitatively measure word-to-word relationships. The specific process may include: pairs of (input words, output words) are obtained using Skip-Gram models or CBOW models (continuous word bag models). And encoding the input word and the output word by using one-hot (single hot coding) to obtain training samples of the model. And finally, carrying the encoded input words and the encoded output words into a neural network for training, wherein the input matrix multiplication input-hidden layer weight matrix result is a word vector result of the input words, so that word embedded vector representation of the history medical data after each word segmentation is obtained. For example, text data in a historic electronic medical record may be extracted as historic medical data, which may include, for example, natural text data and diagnostic field information in an discharge record; further, chinese word segmentation is completed on the text data, and the text data after word segmentation is expressed as X. ICD code field information corresponding to the first page of the medical records can be further extracted and is indicated as Y. By representing historical medical data as word embedded vector representations, the data can be described by distributed vectors, thereby reducing computation and memory.
In step S220, training the machine learning model according to the word embedded vector representation and the real coding information of the historical medical data to obtain the trained machine learning model.
In the embodiment of the disclosure, the true coding information refers to the determined coding information marked manually, and the coding information refers to ICD (International Classification of Diseases, international disease classification) coding. In particular, according to certain characteristics of the disease, the disease is classified according to rules and expressed by a coding method. The coding information of different disease information is different, and each coding information corresponds to one disease information respectively. The encoded information herein refers to encoded information that has been classified according to historical medical data, such as code 1, code 2, and the like.
The machine learning model may be trained from historical medical data, and the actual encoded information may be considered as a label for the historical medical data, e.g., "1" for code 1, "2" for code 2, etc. On the basis, word embedded vector representation corresponding to the segmented historical medical data can be used as input of a machine learning model, and prediction coding information obtained through the machine learning model and the medical attention model is used as output to train the model. In training a machine learning model, a gradient descent algorithm may be employed for model training. The gradient descent algorithm may include batch gradient descent or random gradient descent, etc., and is not particularly limited herein. Of course, other algorithms may be used for model training, and are not particularly limited herein.
A flowchart for training a machine learning model is schematically shown in fig. 3, and the steps in fig. 3 are specific implementations of step S220, and referring to fig. 3, the steps mainly include:
In step S310, predictive coding information for the historical medical data is determined from the word embedding vector representation.
In the embodiment of the disclosure, the predictive coding information refers to the coding information of the historical medical data automatically output through the machine learning model and the medical attention model. The prediction encoding information may be the same as or different from the actual encoding information, and is not limited herein.
A flowchart for determining predictive coding information is schematically shown in fig. 4, and the steps in fig. 4 are specific implementations of step S310, and referring to fig. 4, mainly include steps S410 to S440, where:
In step S410, the word-embedded vector representation is input into a machine learning model to obtain a historical feature representation;
in step S420, a coding vector is configured for a plurality of coding information, and a history feature representation corresponding to the history medical data is matched with the coding vector to determine a matching degree;
In step S430, weighting the historical feature representation according to the matching degree to obtain the text representation;
in step S440, the predictive coding information of the historical medical data is determined from the textual representation.
In the disclosed embodiment, since the machine learning model is trained in conjunction with the medical attention model at this time, the output of the machine learning model is a feature representation and the input thereof is a word embedded vector representation. When the predictive coding information is determined, the predictive coding information is obtained according to the output of the machine learning model, so that word embedded vector representations of the historical medical data can be input into the machine learning model to obtain historical characteristic representations corresponding to the historical medical data, and the predictive coding information of the historical medical data can be accurately predicted according to the historical characteristic representations.
For example, the word embedded vector representation and the true coding information < X, Y > are input into a convolutional neural network, the data characterization is completed by the convolutional neural network to obtain a feature representation, the text representation is completed by the medical attention model, and the model training is further completed by a gradient descent method.
In the embodiment of the disclosure, if the classification result is a label or probability representing the encoded information, the output vector refers to the encoded vector corresponding to the encoded information. Based on this, one code vector may be respectively configured for each code information to be output through the medical attention model for representing the main feature of the code information, the code vector corresponds to each code information one by one, and the code vector corresponding to each code information may be different.
Further, the historical feature representation corresponding to the historical medical data and the encoding vector corresponding to the encoding information can be matched to determine the matching degree between the historical feature representation and the encoding vector, and the historical feature representation is weighted according to the matching degree to obtain the text representation of the historical medical data. The matching degree refers to the attention weight of the calculated historical medical data, and the text representation of the historical medical data can be obtained according to the historical characteristic representation and the attention weight. Specifically, each piece of encoded information will have a coded vector representation, and each tag or word in the historical medical data will be understood as a word, and will also have a vector representation, such as a historical feature representation. The two vectors are identical in length, so that the two vectors can be matched to obtain the matching degree between the two vectors. Specifically, the historical feature representation may be subjected to a first logical operation with the encoding vector to obtain a degree of matching between the historical feature representation and the encoding information. The first logical operation may be a dot product operation for converting two vectors into a degree of match for one scalar representation. The degree of matching may be used to measure the degree of matching between the encoded information and the historical medical data (words). The matching degree can be a score or a numerical value, and is high in matching degree and strong in correlation; the degree of matching is low, indicating weak correlation. By matching words or phrases with high scores, the interpretability or basis of the encoded information can be increased.
Further, after the matching degree is obtained, the historical feature representation of the historical medical data may be weighted based on the matching degree to obtain a textual representation describing the entire historical medical data. The specific process may include: and carrying out second logic operation on the matching degree and each history characteristic representation to determine a text representation corresponding to the history medical data. The second logical operation refers to a multiplication operation as well as an addition operation. Specifically, after the obtained matching degree, a matching score exists for each word in the historical medical data, and all the scores can be normalized first, i.e. the scores are scaled after being mapped according to a certain proportion or according to a certain function, so that the sum of the scaled scores is equal to 1. Further, the matching degree after normalization can be multiplied by the historical feature representation of the words in each historical medical data respectively to obtain a plurality of multiplied results; and adding the multiplied results to obtain a weighted representation of the words contained in the whole historical medical data, wherein the weighted representation is used as a text representation of the whole historical medical data and used for predicting the coding classification of the historical medical data. It should be noted that, since the encoding vector of each encoded information is different, the text representation corresponding to each encoded information that is finally calculated by the medical attention model is also different.
Further, the text representation may be re-entered into the medical classification model, and predictive coding information for the historical medical data may be obtained. Specifically, feature extraction may be performed through a convolution operation to determine a feature representation and a text representation of the historical medical data, and further performing subsequent convolution and normalization processes, and so on, to determine a category label of the historical medical data. Based on this, the prediction probability or the prediction label of which encoded information the input history medical data belongs to, i.e., which encoded information the history medical data belongs to, can be determined by the medical classification model and the medical attention model. For example, when the prediction probability of the historical medical data 1 belonging to the encoded information 1 is greater than a certain threshold, for example, 0.8, the prediction label corresponding to the historical medical data 1 may be regarded as 1. Further, predictive coding information for the historical medical data may be determined based on the predictive labels or the predictive probabilities. The historical characteristic representation is weighted through calculating the matching degree so as to determine the predictive coding information, and the coding information with larger correlation can be determined for the historical medical data, so that the accuracy of determining the predictive coding information is improved.
In step S320, the machine learning model is trained based on the prediction encoding information and the real encoding information, so as to obtain the trained machine learning model.
On the basis, the prediction coding information is compared with the real coding information corresponding to the segmented historical medical data, and when the prediction coding information and the real coding information are different, the parameters of the machine learning model are continuously adjusted until the prediction coding information and the real coding information are the same. And through multiple iterations, obtaining a trained machine learning model with high accuracy, and taking the trained machine learning model as a medical classification model. The prediction coding information determined by the matching degree is compared with the real coding information, so that a more accurate model can be obtained, and the efficiency and accuracy of model training can be improved.
After the trained machine learning model is obtained and used as the medical classification model, the medical data to be processed of the target object can be input into the medical classification model so that the medical classification model can extract the characteristics of the medical data to be processed, and the extracted characteristic data is subjected to convolution operation, so that the characteristic representation, namely the characteristic vector, corresponding to the medical data to be processed is obtained. Further, the feature representation may be input to a medical attention model to facilitate conversion of the feature representation into a more accurate textual representation by the medical attention model. And then, inputting the text representation into the machine learning model again, and carrying out coding classification recognition on the medical data to be processed according to the text representation so as to obtain a classification result corresponding to each medical data to be processed. The medical classification model obtained through matching degree training can accurately determine the classification result. The medical data to be processed can be screened through the medical attention model and the matching degree, so that important data can be screened out, attention is focused on the important data, and the accuracy of data processing is improved. And words or sentences related to the coding information can be obtained from the medical data to be processed, coding basis is given, and the interpretability is increased.
In step S130, the coding information corresponding to the medical data to be processed is determined according to the classification result.
In the embodiment of the disclosure, the classification result may be a prediction label or a prediction probability, which specifically varies according to the classification of the model. Based on this, the identification result of the medical data to be processed can be determined based on the category of the classification result, and the coding information to which the medical data to be processed belongs can be determined.
According to the technical scheme provided by the embodiment of the disclosure, the medical data to be processed can be automatically encoded through the medical classification model and the medical attention model, so that ICD encoding information of the medical data to be processed is obtained. A natural language processing method is introduced to automatically finish medical record coding recommendation on medical record data, so that the problems of time and labor waste of manual coding, coding efficiency improvement and inconsistent coding information are solved, meanwhile, coding basis is given through related phrases or sentences, specifically, coding basis is provided for each coding information by the medical record data and the coding, and the interpretability based on a medical attention model is provided. The recommended coding result and the corresponding basis are directly traversed, the automatic coding process is completed, and convenience is improved.
In an embodiment of the present disclosure, there is also provided an automatic medical record encoding apparatus, referring to fig. 5, the apparatus 500 mainly includes: a data acquisition module 501, a code classification module 502, and a code determination module 503, wherein:
a data acquisition module 501, which may be used to acquire medical data to be processed;
The code classification module 502 may be configured to perform medical record code classification and identification on the medical data to be processed in combination with a medical classification model and a medical attention model, so as to obtain a classification result of the medical data to be processed;
The code determining module 503 may be configured to determine code information corresponding to the medical data to be processed according to the classification result.
In one exemplary embodiment of the present disclosure, the code classification module includes: the characteristic representation acquisition module is used for carrying out characteristic extraction on the medical data to be processed through the medical classification model so as to determine the characteristic representation of the medical data to be processed; the text representation acquisition module is used for carrying out weighting processing on the characteristic representation based on the medical attention model to obtain a text representation of the medical data to be processed; and the classification result determining module is used for determining the classification result according to the text representation of the medical data to be processed.
In an exemplary embodiment of the present disclosure, the apparatus further comprises: the model training module is used for training the machine learning model through the historical medical data and the real coding information of the historical medical data, and taking the trained machine learning model as the medical classification model.
In one exemplary embodiment of the present disclosure, the model training module includes: the word vector determining module is used for obtaining word embedded vector representations corresponding to the historical medical data; and the training control module is used for training the machine learning model according to the word embedded vector representation and the real coding information of the historical medical data so as to obtain the trained machine learning model.
In one exemplary embodiment of the present disclosure, the word vector determination module includes: the data word segmentation module is used for acquiring first type data and second type data as the historical medical data, and performing word segmentation processing on the historical medical data to obtain segmented historical medical data; and the word vector training module is used for training the word embedded vector representation according to the segmented historical medical data.
In one exemplary embodiment of the present disclosure, training the machine learning model based on the word embedded vector representation and the true encoded information of the historical medical data to obtain the trained machine learning model includes: determining predictive coding information for the historical medical data based on the word embedded vector representation; and training the machine learning model based on the predictive coding information and the real coding information to obtain the trained machine learning model.
In one exemplary embodiment of the present disclosure, determining predictive coding information for the historical medical data from the word embedding vector representation includes: inputting the word embedded vector representation into the machine learning model to obtain a historical feature representation; configuring coding vectors for a plurality of pieces of coding information, and matching historical characteristic representations corresponding to the historical medical data with the coding vectors to determine matching degree; weighting the historical characteristic representation according to the matching degree to obtain a text representation; the predictive coding information for the historical medical data is determined from the text representation.
In one exemplary embodiment of the present disclosure, matching the historical feature representation corresponding to the historical medical data with the encoding vector to determine a degree of matching includes: and performing first logic operation on the historical characteristic representation and the coding vector to obtain the matching degree between the historical characteristic representation and the coding information.
In an exemplary embodiment of the present disclosure, weighting the historical feature representation according to the degree of matching includes: and performing a second logic operation on the matching degree and each historical characteristic representation to determine the text representation.
It should be noted that, each functional module of the automatic medical record encoding device in the embodiment of the present disclosure is the same as the steps of the above-mentioned example embodiment of the automatic medical record encoding method, so that a detailed description thereof is omitted herein.
It should be noted that although in the above detailed description several modules or units of a device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit in accordance with embodiments of the invention. Conversely, the features and functions of one module or unit described above may be further divided into a plurality of modules or units to be embodied.
Furthermore, although the steps of the methods of the present invention are depicted in the accompanying drawings in a particular order, this is not required to or suggested that the steps must be performed in this particular order or that all of the steps shown be performed in order to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step to perform, and/or one step decomposed into multiple steps to perform, etc.
In an exemplary embodiment of the present invention, an electronic device capable of implementing the above method is also provided.
Those skilled in the art will appreciate that the various aspects of the invention may be implemented as a system, method, or program product. Accordingly, aspects of the invention may be embodied in the following forms, namely: an entirely hardware embodiment, an entirely software embodiment (including firmware, micro-code, etc.) or an embodiment combining hardware and software aspects may be referred to herein as a "circuit," module "or" system.
An electronic device 600 according to this embodiment of the invention is described below with reference to fig. 6. The electronic device 600 shown in fig. 6 is merely an example and should not be construed to limit the functionality and scope of use of embodiments of the present disclosure in any way.
As shown in fig. 6, the electronic device 600 is in the form of a general purpose computing device. Components of electronic device 600 may include, but are not limited to: the at least one processing unit 610, the at least one memory unit 620, a bus 630 connecting the different system components (including the memory unit 620 and the processing unit 610), and a display unit 640.
Wherein the storage unit stores program code that is executable by the processing unit 610 such that the processing unit 610 performs steps according to various exemplary embodiments of the present invention described in the above-described "exemplary methods" section of the present specification. For example, the processing unit 610 may perform the steps as shown in fig. 1.
The storage unit 620 may include readable media in the form of volatile storage units, such as Random Access Memory (RAM) 6201 and/or cache memory unit 6202, and may further include Read Only Memory (ROM) 6203.
The storage unit 620 may also include a program/utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment.
Bus 630 may be a local bus representing one or more of several types of bus structures including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or using any of a variety of bus architectures.
The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboard, pointing device, bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 600, and/or any device (e.g., router, modem, etc.) that enables the electronic device 600 to communicate with one or more other computing devices. Such communication may occur through an input/output (I/O) interface 650. Also, electronic device 600 may communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network, such as the Internet, through network adapter 660. As shown, network adapter 660 communicates with other modules of electronic device 600 over bus 630. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with electronic device 600, including, but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, data backup storage systems, and the like.
From the above description of embodiments, those skilled in the art will readily appreciate that the example embodiments described herein may be implemented in software, or may be implemented in software in combination with the necessary hardware. Thus, the technical solution according to the embodiment of the present invention may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a usb disk, a mobile hard disk, etc.) or on a network, and includes several instructions to cause a computing device (may be a personal computer, a server, a mobile terminal, or a network device, etc.) to perform the medical records automatic encoding method according to the embodiment of the present invention.
In an exemplary embodiment of the present invention, a computer-readable storage medium having stored thereon a program product capable of implementing the method described above in the present specification is also provided. In some possible embodiments, the various aspects of the invention may also be implemented in the form of a program product comprising program code for causing a terminal device to carry out the steps according to the various exemplary embodiments of the invention as described in the "exemplary methods" section of this specification, when said program product is run on the terminal device.
A program product for implementing the above-described method according to an embodiment of the present invention may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The computer readable signal medium may include a data signal propagated in baseband or as part of a carrier wave with readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A readable signal medium may also be any readable medium that is not a readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).
Furthermore, the above-described drawings are only schematic illustrations of processes included in the method according to the exemplary embodiment of the present invention, and are not intended to be limiting. It will be readily appreciated that the processes shown in the above figures do not indicate or limit the temporal order of these processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, for example, among a plurality of modules.
Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
Claims (8)
1. An automatic medical record coding method, which is characterized by comprising the following steps:
Acquiring medical data to be processed;
Performing medical record coding classification recognition on the medical data to be processed by combining a medical classification model and a medical attention model to obtain a classification result of the medical data to be processed;
Determining coding information corresponding to the medical data to be processed according to the classification result;
the method further comprises the steps of:
Training a machine learning model through historical medical data and real coding information of the historical medical data, and taking the trained machine learning model as the medical classification model;
the training of the machine learning model by the historical medical data and the coding information of the historical medical data comprises the following steps:
acquiring word embedding vector representations corresponding to the historical medical data;
Training the machine learning model according to the word embedded vector representation and the true coding information of the historical medical data to obtain the trained machine learning model, comprising: determining predictive coding information for the historical medical data based on the word embedded vector representation; training the machine learning model based on the predictive coding information and the real coding information to obtain the trained machine learning model; wherein the historical medical data comprises a first type of data and a second type of data, and the sources of the first type of data and the second type of data are different;
The determining predictive coding information for the historical medical data from the word embedded vector representation includes: inputting the word embedded vector representation into the machine learning model to obtain a historical feature representation;
Configuring coding vectors for a plurality of pieces of coding information, and matching historical characteristic representations corresponding to the historical medical data with the coding vectors to determine matching degree;
Weighting the historical characteristic representation according to the matching degree to obtain a text representation, wherein after the obtained matching degree, all scores are subjected to normalization processing; multiplying the normalized matching degree by historical feature representations of words in each historical medical data respectively to obtain a plurality of multiplied results; adding the multiplied results to obtain a weighted representation of words contained in the whole historical medical data to be used as a text representation of the whole historical medical data;
The predictive coding information for the historical medical data is determined from the text representation.
2. The medical records automatic coding method according to claim 1, wherein performing medical records coding classification on the medical data to be processed in combination with a medical classification model and a medical attention model to obtain a classification result of the medical data to be processed comprises:
extracting features of the medical data to be processed through the medical classification model to determine feature representation of the medical data to be processed;
weighting the characteristic representation based on the medical attention model to obtain a text representation of the medical data to be processed;
And determining the classification result according to the text representation of the medical data to be processed.
3. The automated medical records encoding method of claim 1, wherein obtaining the word embedding vector representation corresponding to the historical medical data comprises:
acquiring first type data and second type data as the historical medical data, and performing word segmentation on the historical medical data to obtain segmented historical medical data;
Training the word embedding vector representation according to the segmented historical medical data.
4. The automated medical records encoding method of claim 1, wherein matching the historical feature representation corresponding to the historical medical data to the encoded vector to determine a degree of matching comprises:
and performing first logic operation on the historical characteristic representation and the coding vector to obtain the matching degree between the historical characteristic representation and the coding information.
5. The automated medical records encoding method of claim 1, wherein weighting the historical feature representation according to the degree of matching to obtain a textual representation comprises:
and performing a second logic operation on the matching degree and each historical characteristic representation to determine the text representation.
6. An automatic medical records encoding device, comprising:
the data acquisition module is used for acquiring medical data to be processed;
The coding classification module is used for carrying out medical record coding classification identification on the medical data to be processed by combining the medical classification model and the medical attention model so as to obtain a classification result of the medical data to be processed;
The code determining module is used for determining code information corresponding to the medical data to be processed according to the classification result;
the apparatus may further comprise a device for controlling the operation of the apparatus,
The model training module is used for training a machine learning model through historical medical data and real coding information of the historical medical data, and taking the trained machine learning model as the medical classification model;
The model training module comprises: the word vector determining module is used for obtaining word embedded vector representations corresponding to the historical medical data; the training control module is configured to train the machine learning model according to the word embedded vector representation and the real coding information of the historical medical data, so as to obtain the trained machine learning model, and includes: determining predictive coding information for the historical medical data based on the word embedded vector representation; training the machine learning model based on the predictive coding information and the real coding information to obtain the trained machine learning model, wherein the historical medical data comprises first type data and second type data, and the sources of the first type data and the second type data are different;
The determining predictive coding information for the historical medical data from the word embedded vector representation includes: inputting the word embedded vector representation into the machine learning model to obtain a historical feature representation; configuring coding vectors for a plurality of pieces of coding information, and matching historical characteristic representations corresponding to the historical medical data with the coding vectors to determine matching degree; weighting the historical characteristic representation according to the matching degree to obtain a text representation, wherein after the obtained matching degree, all scores are subjected to normalization processing; multiplying the normalized matching degree by historical feature representations of words in each historical medical data respectively to obtain a plurality of multiplied results; adding the multiplied results to obtain a weighted representation of words contained in the whole historical medical data to be used as a text representation of the whole historical medical data; the predictive coding information for the historical medical data is determined from the text representation.
7. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the method for automatically encoding medical records according to any one of claims 1-5.
8. An electronic device, comprising:
A processor; and
A memory for storing executable instructions of the processor;
Wherein the processor is configured to perform the medical records automatic encoding method of any one of claims 1-5 via execution of the executable instructions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911310848.3A CN113012774B (en) | 2019-12-18 | 2019-12-18 | Automatic medical record coding method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911310848.3A CN113012774B (en) | 2019-12-18 | 2019-12-18 | Automatic medical record coding method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113012774A CN113012774A (en) | 2021-06-22 |
| CN113012774B true CN113012774B (en) | 2024-08-30 |
Family
ID=76380903
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911310848.3A Active CN113012774B (en) | 2019-12-18 | 2019-12-18 | Automatic medical record coding method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113012774B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113627525B (en) * | 2021-08-10 | 2024-06-18 | 工银科技有限公司 | Training method of feature extraction model, medical insurance risk identification method and device |
| CN117423423B (en) * | 2023-12-18 | 2024-02-13 | 四川互慧软件有限公司 | Health record integration method, equipment and medium based on convolutional neural network |
| CN117637092B (en) * | 2024-01-24 | 2024-04-23 | 创智和宇信息技术股份有限公司 | Medical record precoding method and device based on artificial intelligence model |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2009302301B2 (en) * | 2008-10-08 | 2016-06-09 | The Mclean Hospital Corporation | Methods for measuring the magnitude of attention and motor activity disturbance in a subject |
| US10811125B2 (en) * | 2017-08-21 | 2020-10-20 | International Business Machines Corporation | Cognitive framework to identify medical case safety reports in free form text |
| KR102225894B1 (en) * | 2018-04-24 | 2021-03-11 | 네이버 주식회사 | Method and system for predicting disease prognosis from patient medical history using deep attention network |
| CN109471945B (en) * | 2018-11-12 | 2021-11-23 | 中山大学 | Deep learning-based medical text classification method and device and storage medium |
| CN109582955B (en) * | 2018-11-14 | 2023-04-07 | 金色熊猫有限公司 | Method, apparatus and medium for standardizing medical terms |
| CN109697285B (en) * | 2018-12-13 | 2022-06-21 | 中南大学 | Hierarchical BilSt Chinese electronic medical record disease coding and labeling method for enhancing semantic representation |
| CN110490251B (en) * | 2019-03-08 | 2022-07-01 | 腾讯科技（深圳）有限公司 | Artificial intelligence-based prediction classification model obtaining method and device and storage medium |
| CN110491465B (en) * | 2019-08-20 | 2020-09-15 | 山东众阳健康科技集团有限公司 | Disease classification coding method, system, device and medium based on deep learning |
-
2019
- 2019-12-18 CN CN201911310848.3A patent/CN113012774B/en active Active
Non-Patent Citations (1)
| Title |
|---|
| A Multi-channel Convolutional Neural Network for ICD Coding;Jie Chen;2019 IEEE 14th International Conference on Intelligent Systems and Knowledge Engineering (ISKE);第1178-1184页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113012774A (en) | 2021-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11790171B2 (en) | Computer-implemented natural language understanding of medical reports | |
| KR102153920B1 (en) | System and method for interpreting medical images through the generation of refined artificial intelligence reinforcement learning data | |
| CN107729313B (en) | Deep neural network-based polyphone pronunciation distinguishing method and device | |
| CN112183101A (en) | Text intention recognition method and device, electronic equipment and storage medium | |
| CN111564223B (en) | Infectious disease survival probability prediction method, and prediction model training method and device | |
| CN112541056B (en) | Medical term standardization method, device, electronic equipment and storage medium | |
| CN108563624A (en) | A kind of spatial term method based on deep learning | |
| CN113012774B (en) | Automatic medical record coding method and device, electronic equipment and storage medium | |
| CN112489740B (en) | Medical record detection method, training method of related model, related equipment and device | |
| CN112002410A (en) | Infectious disease state prediction method and device, storage medium, and electronic device | |
| CN110223134B (en) | Product recommendation method based on voice recognition and related equipment | |
| CN111079432A (en) | Text detection method and device, electronic equipment and storage medium | |
| CN114298050A (en) | Model training method, entity relation extraction method, device, medium and equipment | |
| CN111145903A (en) | Method and device for acquiring vertigo inquiry text, electronic equipment and inquiry system | |
| CN112364664A (en) | Method and device for training intention recognition model and intention recognition and storage medium | |
| CN111755090A (en) | Medical record searching method, medical record searching device, storage medium and electronic equipment | |
| CN113722507B (en) | Hospitalization cost prediction method and device based on knowledge graph and computer equipment | |
| CN115762721A (en) | Medical image quality control method and system based on computer vision technology | |
| CN111161884A (en) | Disease prediction method, device, equipment and medium for unbalanced data | |
| CN112100360B (en) | Dialogue response method, device and system based on vector retrieval | |
| CN111403028B (en) | Medical text classification method and device, storage medium and electronic equipment | |
| CN118277573A (en) | Pre-hospital emergency text classification labeling method based on ChatGLM model, electronic equipment, storage medium and computer program product | |
| CN112700862A (en) | Target department determining method and device, electronic equipment and storage medium | |
| CN116564539B (en) | Medical similar case recommending method and system based on information extraction and entity normalization | |
| CN111081325A (en) | Medical data processing method and 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |