CN103488752A - POI (point of interest) searching method - Google Patents
POI (point of interest) searching method Download PDFInfo
- Publication number
- CN103488752A CN103488752A CN201310438035.9A CN201310438035A CN103488752A CN 103488752 A CN103488752 A CN 103488752A CN 201310438035 A CN201310438035 A CN 201310438035A CN 103488752 A CN103488752 A CN 103488752A
- Authority
- CN
- China
- Prior art keywords
- retrieval
- result
- user
- input
- searching
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/36—Input/output arrangements for on-board computers
- G01C21/3679—Retrieval, searching and output of POI information, e.g. hotels, restaurants, shops, filling stations, parking facilities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/907—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Library & Information Science (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Disclosed is a POI searching method. The intelligent searching starts by waiting for input of a user, which can be achieved through input method handwriting, pinyin, letter or voice input, audio files, map image files or text files; the unprocessed information can be converted into text symbols instead of unnatural languages via natural language input; when searching results are already provided, a self-learning process of the searching process is performed to modify built-in searching comprehension rules, thinking association trees through the relationship of user searching keywords and actually-executed searching steps; then searching segment backup is performed; the searching process is completed and searching results are provided to the user. The POI searching method has the advantages that a user input analyzing process is taken into consideration, and the whole searching process is completed on the basis of comprehension on user searching objects.
Description
Technical field
The present invention relates to in-vehicle navigation apparatus retrieval technique field, particularly a kind of search method of POI intelligent retrieval.
Background technology
There has been the search function of several comparative maturities at present in retrieval about POI, the user can be according to different search conditions, select different search functions to be retrieved, but the user for first use navigation software, can also manage unclear main threads for the moment, feel to have so many search functions, which this retrieve to look for the place that will go with, is a very scabrous problem.
That merges search function whether can address this problem, although the upper merging of some functions is arranged, if but still rest on the combination on physical property, the effect often also can not reached, because it is complicated that function has become, function search condition before is all different, only accomplishes a real inside search function that permeates, real intelligence is got up, and could fundamentally address this problem.
Summary of the invention
The objective of the invention is to design for addressing the above problem, intelligent retrieval allows the user not go to be concerned about what difference is search function have, be applicable to which type of retrieval scene, the user only need to take notice of the purpose that oneself is retrieved, whether can by key word know express just passable, this use habit relatively meets our daily behavioral trait, and the spy provides a kind of search method of POI intelligent retrieval.
The invention provides a kind of search method of POI intelligent retrieval, it is characterized in that: the search method of described POI intelligent retrieval, considered that the understanding based on to the user search purpose, complete whole retrieving, sees Fig. 1 to the analytic process of user's input:
Intelligent retrieval starts, at first wait for user's input, user's input can be by hand-written, the phonetic of input method, letter, voice typing, audio file, map picture file, text, user's input is not taked to too much restriction here, meet user's use habit as far as possible; For above unprocessed user's input information, need, through the natural language input, these unprocessed information to be changed into to textual character, and unnatural language; For example: the voice signal to input, need speech recognition to change into voice messaging, picture file also needs the text message in picture is changed into to the directivity of expressing in text or text;
Textual character for after transforming, carry out feature extraction according to default rule, carries out the conversion of text, and the replacement of word and symbol, removes some meaningless separation words, selectively removes the word that some punctuation marks and discrimination are not high, natural language understanding is based on the Statistical Probabilistic Models trained, from the pattern rules storehouse, find approximate pattern, approximate mode can have a plurality of, each pattern can be set in a corresponding retrieval thinking association, see Fig. 2, a for example user search " station, Shenyang ", can be according to the input at " station, Shenyang ", search corresponding thinking association tree, if find, can be according to thinking association tree, carry out result for retrieval, at first can retrieve the station, Shenyang as railway station, second step can be retrieved the vehicles of station, Shenyang periphery, the information of subway station can preferentially be provided according to probability, the 3rd step, can retrieve peripheral fast food, the 4th step, can stay by the retrieval periphery, the 5th step, just can retrieve the POI that other comprise " railway station " title, if user's input is " near the company station, Shenyang ", company does not appear in retrieval thinking association tree, by directly, according to built-in retrieval model, is retrieved,
Situation about occurring for a plurality of search modes, can generate a plurality of searching steps, next can carry out polymerization to these searching steps, and to get rid of be obviously irrational searching step, and these processing procedures are that the rule defined by some is carried out; Through several take turns such processing procedure after, can generate some rough machined steps, these steps guarantee it is to exist reasonably;
For these rough machined steps, also need to add preprocessing process, process rear process etc., and, to after each key element of step comprehensive quantification, be optimized sequence, just complete the compile optimization of whole searching step;
After these steps generate, be not the concrete operations that will begin in a minute and retrieve, whether exist but need to first inquire about retrieve fragment, if present, just directly carry out the fragment of retrieval; In retrieve fragment, can comprise a step searching step and a step result for retrieval; Carry out retrieve fragment branch, generally, retrieving only need complete the searching step of vacancy in searching step, and at suitable " breakpoint ", continuing to carry out gets final product; Can effectively accelerate like this process of retrieval, after carrying out retrieve fragment, can give result for retrieval and arrange process;
If there is no retrieve fragment, can generate the searching step inventory, in this process, and the process of meeting initial interrogation and scheduling searching step, and opening space is preserved the contextual information between search argument and searching step etc.; One of the every execution of searching step all can subtract 1 in the queue of searching step inventory, and when the searching step inventory reduces to 0, all result for retrieval have been worked it out in retrieval, can give afterwards result for retrieval and arrange process;
Result for retrieval arranges process and can be marked and classify the result retrieved, and to result for retrieval the sort result according to scoring and classification, then the result of sequence merged, returned to the user with regard to the result that can be used as whole retrieving;
When according to after the searching step inventory is carried out before, when occurring that result for retrieval is non-existent, need consideration before whether user's retrieval understanding content to be had problems, therefore there is one here for the feedback mechanism of understanding error; Feedback mechanism can carry out re-organized to retrieval, and amplification search condition progressively, at first phase out the word that those discriminations are not high, and the word that the reserved area calibration is high, until to cancelling the high word of discrimination, the word that the reserved area calibration is not high; When result having occurred, and while meeting the termination condition of retrieval, just no longer continue to have amplified search condition, processing procedure is can be by premature termination, and then gives result for retrieval and arrange process, and each retrieval all can make great efforts to attempt offering the user search result;
When result for retrieval provides, there is the self study process of a retrieving, utilize the key word of user search and the relation of the actual searching step of carrying out to revise the rule of retrieving built-in understanding, and thinking association tree etc.; And backup retrieve fragment; So far, retrieving finishes, and the result retrieved is offered to the user.
The flow process of intelligent retrieval is as follows:
Outside input: for receiving user's input, the module usually used as direct and user interactions, provide multiple input mode, meets user's use habit, for example: user speech input, user's handwriting input etc.;
Feature extraction: to user's input, no matter be the content of input, or the behavior details of input, comprise the input of symbol, the input of capital and small letter etc., or or the key word input repeatedly to inputting, after being identified as useful feature, all as feature, can be recorded and extract;
Text-converted: the content of feature extraction further need to be changed into to content of text, some feature extraction out is derived from sound, some feature extraction out is derived from picture, and some feature extraction out is derived from string number, and these all need these contents are converted to the text implication of representative; When the process of text-converted, the situation of ambiguity appears if there is the explanation of text, these ambiguities need to be eliminated, arranged discrimination according to the dictionary of training and the rule of word coupling, and result is carried out to participle, part of speech role's mark;
Semantic understanding: this module can in the pattern rules storehouse, be carried out the coupling of pattern rules to the result of text-converted, can generate the main execution step of retrieval;
Retrieve fragment: core is a cache module, the result for retrieval of each step of meeting buffer memory, the context that each step is carried out, the regular correction result of also not preserving and statistical information etc.; And provide unified access interface and retrieve fragment classification storage and merging, swapping in and out strategy etc. of the same type; Allow a retrieving, in the situation that retrieve fragment exists, can dispense some steps, and have the possibility that each fragment is had to secondary processing;
Searching step generates: in the non-existent situation of asked retrieve fragment, searching step that will complete, searching step can be to the step of semantic understanding, carry out the compile optimization processing, after can considering the key elements such as performance, internal memory, generate the searching step of a reasonable set, add the flow process that is connected between pretreated flow process, step and process after flow process etc., will finally produce the inventory of a searching step; Meanwhile, also can complete the initialization context variable and open up relevant memory headroom, for the step process of retrieval is prepared;
Searching step is processed: this process can be according to the content of request, and according to the service logic of retrieval, the data-base content of traversal retrieval, obtain the result for retrieval that meets querying condition; The number of steps defined when the searching step inventory is greater than zero, and searching step is processed and will be called repeatedly; The end of each step, all can preserve the context of retrieval, for next step retrieval and record retrieval fragment, comes;
Result for retrieval is processed: result for retrieval processing meeting by the result of retrieval classified, the operations such as sequence, merging, the result of this step just can be given to external output module, for being exported to external device;
Error Feedback is processed: to after the generation of primary retrieval step, searching step are processed, result for retrieval does not exist, intelligent retrieval system can judge that there is error in the understanding to user's input, need to revise querying condition, the Error Feedback processing can regenerate searching step, carry out new retrieval, after meeting the condition finished, just can stop retrieval;
Adaptive learning: this is a study module, can carry out adaptive study according to the result of retrieval and user's input, reaches the purpose of continuous adaptation user use habit; According to user's repeatedly retrieval, the rule of the continuous update the system acquiescence of meeting, can affect the result of retrieval and the sequence of result for retrieval etc.;
External output: be exactly user interface, offer the result of user search, to the result for retrieval of user's request, replied.
Advantage of the present invention:
The search method of POI intelligent retrieval of the present invention, the embodiment of intelligent retrieval, considered that in this scheme the understanding based on to the user search purpose, complete whole retrieving to the analytic process of user's input.
The accompanying drawing explanation
Below in conjunction with drawings and the embodiments, the present invention is further detailed explanation:
Fig. 1 is intelligent retrieval process schematic diagram;
Fig. 2 is thinking association tree schematic diagram;
Fig. 3 is the intelligent retrieval processing flow chart.
Embodiment
Embodiment 1
The invention provides a kind of search method of POI intelligent retrieval, it is characterized in that: the search method of described POI intelligent retrieval, considered that the understanding based on to the user search purpose, complete whole retrieving, sees Fig. 1 to the analytic process of user's input:
Intelligent retrieval starts, at first wait for user's input, user's input can be by hand-written, the phonetic of input method, letter, voice typing, audio file, map picture file, text, user's input is not taked to too much restriction here, meet user's use habit as far as possible; For above unprocessed user's input information, need, through the natural language input, these unprocessed information to be changed into to textual character, and unnatural language; For example: the voice signal to input, need speech recognition to change into voice messaging, picture file also needs the text message in picture is changed into to the directivity of expressing in text or text;
Textual character for after transforming, carry out feature extraction according to default rule, carries out the conversion of text, and the replacement of word and symbol, removes some meaningless separation words, selectively removes the word that some punctuation marks and discrimination are not high, natural language understanding is based on the Statistical Probabilistic Models trained, from the pattern rules storehouse, find approximate pattern, approximate mode can have a plurality of, each pattern can be set in a corresponding retrieval thinking association, see Fig. 2, a for example user search " station, Shenyang ", can be according to the input at " station, Shenyang ", search corresponding thinking association tree, if find, can be according to thinking association tree, carry out result for retrieval, at first can retrieve the station, Shenyang as railway station, second step can be retrieved the vehicles of station, Shenyang periphery, the information of subway station can preferentially be provided according to probability, the 3rd step, can retrieve peripheral fast food, the 4th step, can stay by the retrieval periphery, the 5th step, just can retrieve the POI that other comprise " railway station " title, if user's input is " near the company station, Shenyang ", company does not appear in retrieval thinking association tree, by directly, according to built-in retrieval model, is retrieved,
Situation about occurring for a plurality of search modes, can generate a plurality of searching steps, next can carry out polymerization to these searching steps, and to get rid of be obviously irrational searching step, and these processing procedures are that the rule defined by some is carried out; Through several take turns such processing procedure after, can generate some rough machined steps, these steps guarantee it is to exist reasonably;
For these rough machined steps, also need to add preprocessing process, process rear process etc., and, to after each key element of step comprehensive quantification, be optimized sequence, just complete the compile optimization of whole searching step;
After these steps generate, be not the concrete operations that will begin in a minute and retrieve, whether exist but need to first inquire about retrieve fragment, if present, just directly carry out the fragment of retrieval; In retrieve fragment, can comprise a step searching step and a step result for retrieval; Carry out retrieve fragment branch, generally, retrieving only need complete the searching step of vacancy in searching step, and at suitable " breakpoint ", continuing to carry out gets final product; Can effectively accelerate like this process of retrieval, after carrying out retrieve fragment, can give result for retrieval and arrange process;
If there is no retrieve fragment, can generate the searching step inventory, in this process, and the process of meeting initial interrogation and scheduling searching step, and opening space is preserved the contextual information between search argument and searching step etc.; One of the every execution of searching step all can subtract 1 in the queue of searching step inventory, and when the searching step inventory reduces to 0, all result for retrieval have been worked it out in retrieval, can give afterwards result for retrieval and arrange process;
Result for retrieval arranges process and can be marked and classify the result retrieved, and to result for retrieval the sort result according to scoring and classification, then the result of sequence merged, returned to the user with regard to the result that can be used as whole retrieving;
When according to after the searching step inventory is carried out before, when occurring that result for retrieval is non-existent, need consideration before whether user's retrieval understanding content to be had problems, therefore there is one here for the feedback mechanism of understanding error; Feedback mechanism can carry out re-organized to retrieval, and amplification search condition progressively, at first phase out the word that those discriminations are not high, and the word that the reserved area calibration is high, until to cancelling the high word of discrimination, the word that the reserved area calibration is not high; When result having occurred, and while meeting the termination condition of retrieval, just no longer continue to have amplified search condition, processing procedure is can be by premature termination, and then gives result for retrieval and arrange process, and each retrieval all can make great efforts to attempt offering the user search result;
When result for retrieval provides, there is the self study process of a retrieving, utilize the key word of user search and the relation of the actual searching step of carrying out to revise the rule of retrieving built-in understanding, and thinking association tree etc.; And backup retrieve fragment; So far, retrieving finishes, and the result retrieved is offered to the user.
The flow process of intelligent retrieval is as follows:
Outside input: for receiving user's input, the module usually used as direct and user interactions, provide multiple input mode, meets user's use habit, for example: user speech input, user's handwriting input etc.;
Feature extraction: to user's input, no matter be the content of input, or the behavior details of input, comprise the input of symbol, the input of capital and small letter etc., or or the key word input repeatedly to inputting, after being identified as useful feature, all as feature, can be recorded and extract;
Text-converted: the content of feature extraction further need to be changed into to content of text, some feature extraction out is derived from sound, some feature extraction out is derived from picture, and some feature extraction out is derived from string number, and these all need these contents are converted to the text implication of representative; When the process of text-converted, the situation of ambiguity appears if there is the explanation of text, these ambiguities need to be eliminated, arranged discrimination according to the dictionary of training and the rule of word coupling, and result is carried out to participle, part of speech role's mark;
Semantic understanding: this module can in the pattern rules storehouse, be carried out the coupling of pattern rules to the result of text-converted, can generate the main execution step of retrieval;
Retrieve fragment: core is a cache module, the result for retrieval of each step of meeting buffer memory, the context that each step is carried out, the regular correction result of also not preserving and statistical information etc.; And provide unified access interface and retrieve fragment classification storage and merging, swapping in and out strategy etc. of the same type; Allow a retrieving, in the situation that retrieve fragment exists, can dispense some steps, and have the possibility that each fragment is had to secondary processing;
Searching step generates: in the non-existent situation of asked retrieve fragment, searching step that will complete, searching step can be to the step of semantic understanding, carry out the compile optimization processing, after can considering the key elements such as performance, internal memory, generate the searching step of a reasonable set, add the flow process that is connected between pretreated flow process, step and process after flow process etc., will finally produce the inventory of a searching step; Meanwhile, also can complete the initialization context variable and open up relevant memory headroom, for the step process of retrieval is prepared;
Searching step is processed: this process can be according to the content of request, and according to the service logic of retrieval, the data-base content of traversal retrieval, obtain the result for retrieval that meets querying condition; The number of steps defined when the searching step inventory is greater than zero, and searching step is processed and will be called repeatedly; The end of each step, all can preserve the context of retrieval, for next step retrieval and record retrieval fragment, comes;
Result for retrieval is processed: result for retrieval processing meeting by the result of retrieval classified, the operations such as sequence, merging, the result of this step just can be given to external output module, for being exported to external device;
Error Feedback is processed: to after the generation of primary retrieval step, searching step are processed, result for retrieval does not exist, intelligent retrieval system can judge that there is error in the understanding to user's input, need to revise querying condition, the Error Feedback processing can regenerate searching step, carry out new retrieval, after meeting the condition finished, just can stop retrieval;
Adaptive learning: this is a study module, can carry out adaptive study according to the result of retrieval and user's input, reaches the purpose of continuous adaptation user use habit; According to user's repeatedly retrieval, the rule of the continuous update the system acquiescence of meeting, can affect the result of retrieval and the sequence of result for retrieval etc.;
External output: be exactly user interface, offer the result of user search, to the result for retrieval of user's request, replied.
Claims (2)
1. the search method of a POI intelligent retrieval, it is characterized in that: the search method of described POI intelligent retrieval, the beginning of intelligent retrieval, at first wait for user's input, user's input can be by hand-written, the phonetic of input method, letter, voice typing, audio file, map picture file, text, do not take too much restriction to user's input here, meets user's use habit as far as possible; For above unprocessed user's input information, need, through the natural language input, these unprocessed information to be changed into to textual character, and unnatural language;
Textual character for after transforming, carry out feature extraction according to default rule, carries out the conversion of text, and the replacement of word and symbol, removes some meaningless separation words, selectively removes the word that some punctuation marks and discrimination are not high; Natural language understanding is based on the Statistical Probabilistic Models trained, and from the pattern rules storehouse, finds approximate pattern, and approximate mode can have a plurality of, and each pattern can be set in a corresponding retrieval thinking association;
Situation about occurring for a plurality of search modes, can generate a plurality of searching steps, next can carry out polymerization to these searching steps, and to get rid of be obviously irrational searching step, and these processing procedures are that the rule defined by some is carried out; Through several take turns such processing procedure after, can generate some rough machined steps, these steps guarantee it is to exist reasonably;
For these rough machined steps, also need to add preprocessing process, process rear process etc., and, to after each key element of step comprehensive quantification, be optimized sequence, just complete the compile optimization of whole searching step;
After these steps generate, be not the concrete operations that will begin in a minute and retrieve, whether exist but need to first inquire about retrieve fragment, if present, just directly carry out the fragment of retrieval; In retrieve fragment, can comprise a step searching step and a step result for retrieval; Carry out retrieve fragment branch, generally, retrieving only need complete the searching step of vacancy in searching step, and at suitable " breakpoint ", continuing to carry out gets final product; Can effectively accelerate like this process of retrieval, after carrying out retrieve fragment, can give result for retrieval and arrange process;
If there is no retrieve fragment, can generate the searching step inventory, in this process, and the process of meeting initial interrogation and scheduling searching step, and opening space is preserved the contextual information between search argument and searching step etc.; One of the every execution of searching step all can subtract 1 in the queue of searching step inventory, and when the searching step inventory reduces to 0, all result for retrieval have been worked it out in retrieval, can give afterwards result for retrieval and arrange process;
Result for retrieval arranges process and can be marked and classify the result retrieved, and to result for retrieval the sort result according to scoring and classification, then the result of sequence merged, returned to the user with regard to the result that can be used as whole retrieving;
When according to after the searching step inventory is carried out before, when occurring that result for retrieval is non-existent, need consideration before whether user's retrieval understanding content to be had problems, therefore there is one here for the feedback mechanism of understanding error; Feedback mechanism can carry out re-organized to retrieval, and amplification search condition progressively, at first phase out the word that those discriminations are not high, and the word that the reserved area calibration is high, until to cancelling the high word of discrimination, the word that the reserved area calibration is not high; When result having occurred, and while meeting the termination condition of retrieval, just no longer continue to have amplified search condition, processing procedure is can be by premature termination, and then gives result for retrieval and arrange process, and each retrieval all can make great efforts to attempt offering the user search result;
When result for retrieval provides, there is the self study process of a retrieving, utilize the key word of user search and the relation of the actual searching step of carrying out to revise the rule of retrieving built-in understanding, and thinking association tree etc.; And backup retrieve fragment; So far, retrieving finishes, and the result retrieved is offered to the user.
2. according to the search method of POI intelligent retrieval claimed in claim 1, it is characterized in that: the flow process of intelligent retrieval is as follows:
Outside input: for receiving user's input, the module usually used as direct and user interactions, provide multiple input mode, meets user's use habit, for example: user speech input, user's handwriting input etc.;
Feature extraction: to user's input, no matter be the content of input, or the behavior details of input, comprise the input of symbol, the input of capital and small letter etc., or or the key word input repeatedly to inputting, after being identified as useful feature, all as feature, can be recorded and extract;
Text-converted: the content of feature extraction further need to be changed into to content of text, some feature extraction out is derived from sound, some feature extraction out is derived from picture, and some feature extraction out is derived from string number, and these all need these contents are converted to the text implication of representative; When the process of text-converted, the situation of ambiguity appears if there is the explanation of text, these ambiguities need to be eliminated, arranged discrimination according to the dictionary of training and the rule of word coupling, and result is carried out to participle, part of speech role's mark;
Semantic understanding: this module can in the pattern rules storehouse, be carried out the coupling of pattern rules to the result of text-converted, can generate the main execution step of retrieval;
Retrieve fragment: core is a cache module, the result for retrieval of each step of meeting buffer memory, the context that each step is carried out, the regular correction result of also not preserving and statistical information etc.; And provide unified access interface and retrieve fragment classification storage and merging, swapping in and out strategy etc. of the same type; Allow a retrieving, in the situation that retrieve fragment exists, can dispense some steps, and have the possibility that each fragment is had to secondary processing;
Searching step generates: in the non-existent situation of asked retrieve fragment, searching step that will complete, searching step can be to the step of semantic understanding, carry out the compile optimization processing, after can considering the key elements such as performance, internal memory, generate the searching step of a reasonable set, add the flow process that is connected between pretreated flow process, step and process after flow process etc., will finally produce the inventory of a searching step; Meanwhile, also can complete the initialization context variable and open up relevant memory headroom, for the step process of retrieval is prepared;
Searching step is processed: this process can be according to the content of request, and according to the service logic of retrieval, the data-base content of traversal retrieval, obtain the result for retrieval that meets querying condition; The number of steps defined when the searching step inventory is greater than zero, and searching step is processed and will be called repeatedly; The end of each step, all can preserve the context of retrieval, for next step retrieval and record retrieval fragment, comes;
Result for retrieval is processed: result for retrieval processing meeting by the result of retrieval classified, the operations such as sequence, merging, the result of this step just can be given to external output module, for being exported to external device;
Error Feedback is processed: to after the generation of primary retrieval step, searching step are processed, result for retrieval does not exist, intelligent retrieval system can judge that there is error in the understanding to user's input, need to revise querying condition, the Error Feedback processing can regenerate searching step, carry out new retrieval, after meeting the condition finished, just can stop retrieval;
Adaptive learning: this is a study module, can carry out adaptive study according to the result of retrieval and user's input, reaches the purpose of continuous adaptation user use habit; According to user's repeatedly retrieval, the rule of the continuous update the system acquiescence of meeting, can affect the result of retrieval and the sequence of result for retrieval etc.;
External output: be exactly user interface, offer the result of user search, to the result for retrieval of user's request, replied.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310438035.9A CN103488752B (en) | 2013-09-24 | 2013-09-24 | A kind of search method of POI intelligent retrievals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310438035.9A CN103488752B (en) | 2013-09-24 | 2013-09-24 | A kind of search method of POI intelligent retrievals |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103488752A true CN103488752A (en) | 2014-01-01 |
| CN103488752B CN103488752B (en) | 2017-08-25 |
Family
ID=49828978
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310438035.9A Active CN103488752B (en) | 2013-09-24 | 2013-09-24 | A kind of search method of POI intelligent retrievals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103488752B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105468468A (en) * | 2015-12-02 | 2016-04-06 | 北京光年无限科技有限公司 | Data error correction method and apparatus facing question answering system |
| CN106057200A (en) * | 2016-06-23 | 2016-10-26 | 广州亿程交通信息有限公司 | Semantic-based interaction system and interaction method |
| CN106462602A (en) * | 2014-05-27 | 2017-02-22 | 爱信艾达株式会社 | Facility output system, facility output method, and facility output program |
| CN106537385A (en) * | 2014-07-16 | 2017-03-22 | 微软技术许可有限责任公司 | Observation-based query interpretation model modification |
| CN106776981A (en) * | 2016-12-06 | 2017-05-31 | 广州市科恩电脑有限公司 | A kind of intelligent search method based on Heuristics |
| CN107292302A (en) * | 2016-03-31 | 2017-10-24 | 高德信息技术有限公司 | Detect the method and system of point of interest in picture |
| CN109344342A (en) * | 2018-12-17 | 2019-02-15 | 北京百度网讯科技有限公司 | Map data retrieval method, apparatus, retrieval server and system |
| CN110516094A (en) * | 2019-08-29 | 2019-11-29 | 百度在线网络技术（北京）有限公司 | De-weight method, device, electronic equipment and the storage medium of class interest point data |
| CN112784088A (en) * | 2019-11-04 | 2021-05-11 | 北京旷视科技有限公司 | Personnel retrieval method, device, electronic equipment and readable storage medium |
| CN113434067A (en) * | 2015-10-01 | 2021-09-24 | 谷歌有限责任公司 | Action suggestions for user-selected content |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1335574A (en) * | 2001-09-05 | 2002-02-13 | 罗笑南 | Intelligent semantic searching method |
| CN101685448A (en) * | 2008-09-28 | 2010-03-31 | 国际商业机器公司 | Method and device for establishing association between query operation of user and search result |
| CN101916294A (en) * | 2010-08-27 | 2010-12-15 | 黄斌 | Method for realizing exact search by utilizing semantic analysis |
| CN102436448A (en) * | 2010-09-29 | 2012-05-02 | 腾讯科技（深圳）有限公司 | Search method and system |
| CN102543082A (en) * | 2012-01-19 | 2012-07-04 | 北京赛德斯汽车信息技术有限公司 | Voice operation method for in-vehicle information service system adopting natural language and voice operation system |
| CN102831177A (en) * | 2012-07-31 | 2012-12-19 | 聚熵信息技术（上海）有限公司 | Statement error correction method and system |
| CN102890689A (en) * | 2011-07-22 | 2013-01-23 | 北京百度网讯科技有限公司 | Method and system for building user interest model |
-
2013
- 2013-09-24 CN CN201310438035.9A patent/CN103488752B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1335574A (en) * | 2001-09-05 | 2002-02-13 | 罗笑南 | Intelligent semantic searching method |
| CN101685448A (en) * | 2008-09-28 | 2010-03-31 | 国际商业机器公司 | Method and device for establishing association between query operation of user and search result |
| CN101916294A (en) * | 2010-08-27 | 2010-12-15 | 黄斌 | Method for realizing exact search by utilizing semantic analysis |
| CN102436448A (en) * | 2010-09-29 | 2012-05-02 | 腾讯科技（深圳）有限公司 | Search method and system |
| CN102890689A (en) * | 2011-07-22 | 2013-01-23 | 北京百度网讯科技有限公司 | Method and system for building user interest model |
| CN102543082A (en) * | 2012-01-19 | 2012-07-04 | 北京赛德斯汽车信息技术有限公司 | Voice operation method for in-vehicle information service system adopting natural language and voice operation system |
| CN102831177A (en) * | 2012-07-31 | 2012-12-19 | 聚熵信息技术（上海）有限公司 | Statement error correction method and system |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106462602A (en) * | 2014-05-27 | 2017-02-22 | 爱信艾达株式会社 | Facility output system, facility output method, and facility output program |
| CN106462602B (en) * | 2014-05-27 | 2019-09-10 | 爱信艾达株式会社 | Facility output system, facility output method and medium |
| CN106537385A (en) * | 2014-07-16 | 2017-03-22 | 微软技术许可有限责任公司 | Observation-based query interpretation model modification |
| CN106537385B (en) * | 2014-07-16 | 2019-07-26 | 微软技术许可有限责任公司 | Inquiry interpretation model modification based on observation |
| US10817554B2 (en) | 2014-07-16 | 2020-10-27 | Microsoft Technology Licensing, Llc | Observation-based query interpretation model modification |
| CN113434067A (en) * | 2015-10-01 | 2021-09-24 | 谷歌有限责任公司 | Action suggestions for user-selected content |
| US12026593B2 (en) | 2015-10-01 | 2024-07-02 | Google Llc | Action suggestions for user-selected content |
| CN105468468A (en) * | 2015-12-02 | 2016-04-06 | 北京光年无限科技有限公司 | Data error correction method and apparatus facing question answering system |
| CN105468468B (en) * | 2015-12-02 | 2018-07-27 | 北京光年无限科技有限公司 | Data error-correcting method towards question answering system and device |
| CN107292302A (en) * | 2016-03-31 | 2017-10-24 | 高德信息技术有限公司 | Detect the method and system of point of interest in picture |
| CN106057200A (en) * | 2016-06-23 | 2016-10-26 | 广州亿程交通信息有限公司 | Semantic-based interaction system and interaction method |
| CN106776981A (en) * | 2016-12-06 | 2017-05-31 | 广州市科恩电脑有限公司 | A kind of intelligent search method based on Heuristics |
| CN109344342A (en) * | 2018-12-17 | 2019-02-15 | 北京百度网讯科技有限公司 | Map data retrieval method, apparatus, retrieval server and system |
| CN110516094A (en) * | 2019-08-29 | 2019-11-29 | 百度在线网络技术（北京）有限公司 | De-weight method, device, electronic equipment and the storage medium of class interest point data |
| CN112784088A (en) * | 2019-11-04 | 2021-05-11 | 北京旷视科技有限公司 | Personnel retrieval method, device, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103488752B (en) | 2017-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103488752A (en) | POI (point of interest) searching method | |
| CN112100349B (en) | Multi-round dialogue method and device, electronic equipment and storage medium | |
| CN106156365B (en) | A kind of generation method and device of knowledge mapping | |
| KR102313028B1 (en) | System and method for voice recognition | |
| CN102479191B (en) | Method and device for providing multi-granularity word segmentation result | |
| US10997223B1 (en) | Subject-specific data set for named entity resolution | |
| US10896222B1 (en) | Subject-specific data set for named entity resolution | |
| WO2018028077A1 (en) | Deep learning based method and device for chinese semantics analysis | |
| CN110276023B (en) | POI transition event discovery method, device, computing equipment and medium | |
| US12032915B2 (en) | Creating and interacting with data records having semantic vectors and natural language expressions produced by a machine-trained model | |
| CN108447471A (en) | Audio recognition method and speech recognition equipment | |
| CN111445898B (en) | Language identification method and device, electronic equipment and storage medium | |
| CN103678684A (en) | Chinese word segmentation method based on navigation information retrieval | |
| CN105917327A (en) | System and method for inputting text into electronic devices | |
| CN111159385A (en) | Template-free universal intelligent question-answering method based on dynamic knowledge graph | |
| US20180173694A1 (en) | Methods and computer systems for named entity verification, named entity verification model training, and phrase expansion | |
| CN111742322A (en) | System and method for domain and language independent definition extraction using deep neural networks | |
| CN101561725B (en) | Method and system of fast handwriting input | |
| CN109933671A (en) | Construct method, apparatus, computer equipment and the storage medium of personal knowledge map | |
| US20230054726A1 (en) | Query-focused extractive text summarization of textual data | |
| CN112287100A (en) | Text recognition method, spelling error correction method and voice recognition method | |
| CN112579733A (en) | Rule matching method, rule matching device, storage medium and electronic equipment | |
| CN112417093B (en) | Model training method and device | |
| CN103631784B (en) | Page content retrieval method and system | |
| CN103377224A (en) | Method and device for recognizing problem types and method and device for establishing recognition models |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address |
Address after: 110000 No. 861-6, shangshengou village, Hunnan District, Shenyang City, Liaoning Province Patentee after: Shenyang Meihang Technology Co.,Ltd. Address before: 110179 13th floor, Debao building, No.1 Jinhui street, Hunnan New District, Shenyang City, Liaoning Province Patentee before: SHENYANG MXNAVI Co.,Ltd. |
|
| CP03 | Change of name, title or address |