CN1908935A - Search method and system of a natural language - Google Patents

Search method and system of a natural language Download PDF

Info

Publication number
CN1908935A
CN1908935A CN 200610104093 CN200610104093A CN1908935A CN 1908935 A CN1908935 A CN 1908935A CN 200610104093 CN200610104093 CN 200610104093 CN 200610104093 A CN200610104093 A CN 200610104093A CN 1908935 A CN1908935 A CN 1908935A
Authority
CN
China
Prior art keywords
natural language
template
statement
tolerant
fault
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
Application number
CN 200610104093
Other languages
Chinese (zh)
Other versions
CN100489854C (en
Inventor
张涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to CNB2006101040938A priority Critical patent/CN100489854C/en
Publication of CN1908935A publication Critical patent/CN1908935A/en
Application granted granted Critical
Publication of CN100489854C publication Critical patent/CN100489854C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The disclosed searching method for natural language comprises: setting natural language template unit included more than one template with every one corresponding to one regular expression; to input language, applying the template and corresponding expression to match and filter the statement for metadata; searching database to obtain output business information. This invention has strong function for natural language learning.

Description

A kind of searching method of natural language and system
Technical field
The present invention relates to the search technique field, particularly about a kind of searching method and system of natural language.
Background technology
The Natural Language Search technology is the product that natural language understanding combines with search technique in fact.But the most of Natural Language Searchs in the existing practical application only are that search input word is carried out participle based on dictionary, and then with the keyword that participle comes out index file are searched for.
Participle mainly is the language at non-English family of languages class, such as Chinese." how beautiful flower for example! " in short, can carry out semantic folding branch to sentence according to Chinese dictionary, promptly participle becomes: beautiful, flower, the main body that these are semantic just also is the keyword that Natural Language Search needs.Because for search, searched content has been carried out the establishment of index in advance by the entry in the dictionary.So, when search, may just can find the content that will search for the mode of reverse find.
But the key of Natural Language Search is to separate at the natural language language.Just how from input, to find out key word information.
Introduce method how to find out key word information in two kinds of prior aries below respectively.
One, based on the reasoning of replying of theme forest:
At specific business, determine the theme that its subject tree is related, by repeatedly mutual, finally by the subject information reasoning and search out the desired information of user with the user.
Flight Information as shown in Figure 1, system is by inquiring each theme of user: date, the city of setting out, arrival city, type etc. again by these information removal search, return to the flight that the user determines.
This scheme also can be used in the CRBT search that inserts based on SMS (short message service), if the bell sound is a song so, the theme of this business just has that the bell musical alphabet claims, the singer of bell sound, may also have lyrics frag info.
The shortcoming of prior art one is:
1) mutual too much inferior, user experience is bad;
2) too tight to the user search subject requirement, only provide the singer just can carry out as music searching;
3) system deployment complexity requires too high to the maintainer;
4) System Fault Tolerance insufficiency of function.
Two, natural language participle search
This method is taken as the leading factor with search technique, and semantic analysis is carried out in user's search input, cooperates part-of-speech tagging, finds out the significant keyword of search engine.And then the index file of business datum retrieved.
As input in music searching " beautiful my family of careless garden ", the here search plan is just used " beauty " so, " careless garden " and " family " searched for.The result well imagines, causes user experience very poor.
The shortcoming of prior art two is:
1) do not understand user's semanteme;
2) accuracy is low.
In sum, also there is not a kind of searching method of natural language can search out the result accurately, easily in the prior art.But also do not have fault tolerant mechanism, in case user's input error can't provide the result.
Summary of the invention
In view of this, fundamental purpose of the present invention provides a kind of searching method and system of natural language, in order to solve in the prior art the inaccurate problem of the Search Results of natural language.
The searching method of a kind of natural language provided by the invention comprises:
A. be a traffic set natural language modular unit, described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template;
B. at the statement of Natural Language Search input, utilize the regular expression of natural language template correspondence that described statement is mated filtration, obtain metadata;
C. the business information of utilizing the metasearch Service Database to obtain exporting.
Each natural language template described in the steps A comprises template and main body template end to end, and described regular expression comprises two parts, and each is template and the main body template part in the corresponding described regular expression respectively end to end.
Step B comprises:
B11, select a template end to end, utilize this end to end the template correspondence regular expression and end to end template described statement is mated filtration, if coupling is filtered successfully, obtain the body matter of this statement, execution in step B13, otherwise, execution in step B12;
B12, judge end to end template whether all coupling finish, if do not have, execution in step B11 then, otherwise, process ends;
B13, main body template of selection utilize the regular expression of each main body template correspondence, main body template that described body matter is mated filtration, if coupling is filtered successfully, filtered data is a metadata, execution in step C, otherwise, execution in step B14;
B14, judge the main body template whether all coupling finish, if do not have, execution in step B13 then, otherwise, process ends.
Step B comprises:
B20, with described statement as the keyword lookup Service Database, judge whether to find, if, execution in step C then, otherwise, execution in step B21;
B21, utilize pre-configured template end to end to this statement is mated filtration, if the match is successful, execution in step B22 then, otherwise, execution in step B23;
B22, the main body statement that obtains after will filtering judge whether to find as the keyword lookup Service Database, if, execution in step C then, otherwise, execution in step B23;
B23, utilize pre-configured main body template that the main body statement is mated filtration, if the match is successful, execution in step C then, otherwise, this statement is accurately searched process ends.
Among the step B11 end to end template be to select at random in the template of never using end to end or Information Selection according to priority;
The main body template is selection or according to priority Information Selection at random in the main body template of never using among the step B13.
Described priority is provided with according to character length, the priority height that character is long.
This method further comprises: set in advance fault-tolerant statement, then step C comprises:
Judge whether contain metadata corresponding service information in the Service Database, if this business information of direct acquisition is arranged, otherwise it is fault-tolerant to utilize fault-tolerant statement that metadata is carried out, and utilizes the metasearch Service Database after fault-tolerant, the business information that obtains exporting.
Described fault-tolerant statement is according to phonetic or semantic the setting; Or described fault-tolerant statement is to recently determining according to the degree of correlation of phonetic and fault-tolerant front and back.
The number of described fault-tolerant statement is one or more.
When the number of fault-tolerant statement is one when above, be to select a fault-tolerant statement to carry out fault-tolerant at random among the step C to metadata;
When the number of fault-tolerant statement is one when above, is that to utilize more than one fault-tolerant statement to carry out fault-tolerant among the step C, and exports each metadata corresponding service data after fault-tolerant.
The search system of a kind of natural language provided by the invention comprises:
The natural language filtration unit is used to preserve metadata corresponding service information, preserves professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template; The regular expression that utilizes natural language template correspondence mates to filter to the statement from the natural language R-T unit and obtains metadata, and metadata corresponding service information is exported to the natural language R-T unit;
The natural language R-T unit is used for the statement of external node equipment is sent to the natural language filtration unit, will export to outside node device from the business information of natural language filtration unit.
Described natural language filtration unit comprises:
Natural language template stores device is used to preserve professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template;
The business datum memory storage is used to preserve metadata corresponding service information;
The natural language matching control apparatus, be used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage, statement from the natural language R-T unit is mated filtration obtain metadata, and the business information that metasearch business datum memory storage obtains is exported to the natural language R-T unit.
Also comprise in the described natural language matching control apparatus:
The natural language matching unit is used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage, the statement from the natural language R-T unit is mated to filter obtain metadata, and metadata is exported to fault-tolerant device;
Fault-tolerant device, be used for searching the business datum memory storage and whether contain metadata corresponding service information, if have, directly this business information is exported to the natural language R-T unit, otherwise, it is fault-tolerant to utilize the fault-tolerant statement of self preserving that metadata is carried out, and utilizes the metasearch business datum memory storage after fault-tolerant to obtain corresponding service information, again this business information is exported to the natural language R-T unit.
The present invention is directed to concrete searching service, adopt the mode of language template to come the pattern match natural language, and adopt phonetic transcriptions of Chinese characters and the degree of correlation is fault-tolerant and synonym is fault-tolerant method, obtain metadata, thereby correct Search Results is provided.
Based on the technical scheme of the invention described above, the present invention has following advantage and characteristics:
1) make search system can discern the input of user's natural language;
2) search system has had fault-tolerant, association function;
3) search system has the ability for natural language learning.
Description of drawings
Fig. 1 in the prior art based on the application example synoptic diagram of replying reasoning of theme forest;
Fig. 2 is for realizing the schematic flow sheet of the inventive method;
Fig. 3 for utilize among the present invention the natural language template with and corresponding regular expression described statement mated filter the idiographic flow synoptic diagram that obtains metadata;
Fig. 4 is applied in the embodiment schematic flow sheet of CRBT Natural Language Search for the present invention;
Fig. 5 is for realizing the structural representation of system of the present invention;
Fig. 6 is the synoptic diagram of modular unit end to end;
Fig. 7 is the synoptic diagram of main body modular unit;
Fig. 8 is the wrong schematic representation of apparatus of bell likeness of the deceased;
Fig. 9 is the fault-tolerant device synoptic diagram of singer's " synonym ".
Embodiment
The present invention is directed to concrete searching service, adopt the regular expression of natural language template and this template correspondence to mate the statement of Natural Language Search input, thereby find out metadata.Metadata is meant and the professional relevant basic data information of a certain class that as the title of the music in the music download business, the singer is exactly a metadata of describing music.Regular expression is pattern or a template of describing a string assemble.Which type of character string is this pattern determined belong to this set, and it is made up of text character and metacharacter.Metacharacter is the character of special implication.
Referring to shown in Figure 2, realize that method of the present invention may further comprise the steps:
Step 201: be a traffic set natural language modular unit, described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template.
As: the natural language modular unit in the CRBT search engine comprises modular unit and main body modular unit end to end.Each end to end modular unit comprise a plurality of templates end to end, main body template unit pack is drawn together a plurality of main body modular units.Wherein, but template and main body template arbitrary combination end to end, thus discern more Natural Language Search input.Modular unit and main body modular unit are as shown in Figure 6 and Figure 7 end to end.
If the statement of search input is " I think that the lustily water that Liu Dehua sings is set to CRBT ".Corresponding template is: " I think handle ... be set to CRBT " and " # that * sings ".Wherein, the former is template end to end, and the latter is the main body template, and special symbol " * " and " # " represent the singer and the song title of bell sound music respectively.The regular expression of natural language template correspondence is: " the ^ head+([ s S]+)+body $ ".
Step 202: the statement at the Natural Language Search input, utilize the regular expression of natural language template correspondence that described statement is mated filtration, obtain metadata.
Step 203: the business information of utilizing the metasearch Service Database to obtain exporting.
The natural language template can comprise template and main body template end to end, and described regular expression comprises two parts, and template and the main body template part of corresponding regular expression respectively end to end respectively.Utilize regular expression just the head and tail of the input of the Natural Language Search on the pattern match can be removed, and then main part is carried out the pattern match of main body template.For the more template of matching times, system is the process that an Automatic Optimal is arranged, thereby improves search performance according to statistical law.
Referring to shown in Figure 3, the detailed process that obtains metadata in the step 202 is as follows:
Step 301: select a template end to end, utilize this end to end the regular expression of template correspondence described statement is mated filtration, if the coupling filter successfully, obtain the body matter of this statement, execution in step 303, otherwise, execution in step 302.
Step 302: judge end to end template whether all coupling finish, if do not have, then execution in step 301 otherwise, process ends.Here, select at random in the template end to end that can never use or Information Selection template end to end according to priority.And, can priority be set according to character length, the priority height that character is long.
Step 303: select a main body template, utilize each main body template correspondence regular expression, described body matter is mated filtration, if the coupling filter successfully, filtered data is a metadata, otherwise, execution in step 304.Here, selection or according to priority Information Selection main body template at random in the main body template that can never use.
Step 304: judge the main body template whether all coupling finish, if do not have, then execution in step 303, otherwise, process ends.
From said process as can be seen, main body is being carried out in the process of pattern-recognition, system can get up " body matter ", " main body template " and " the corresponding regular expression of main body template " triplicity, accurately analyzes metadata wherein.
For example:
The people of the favorite of body matter=Zhang Huimei and Zhang Yusheng hinders that I am the darkest
The # of main body template=* and *
Then system can find out the singer and is: Zhang Huimei, Zhang Yusheng; Song is by name: " people of favorite hinder me the darkest " and the singer can not occur is not: the favorite of Zhang Huishu and Zhang Yusheng; Song is by name: " people hinder me the darkest ".
In addition, the main body template will be followed template and grow earlier the principle of afterwards lacking in the process of carrying out pattern match.For example " #* that * sings " will mate before at " # of * ", otherwise the metadata that parses is inaccurate.
In order to improve user's availability, the present invention also provides a kind of fault tolerant mechanism.This need set in advance fault-tolerant statement, when in step 203, searching for, need at first judge and whether contain metadata corresponding service information in the Service Database, if this business information of direct acquisition is arranged, otherwise, it is fault-tolerant to utilize fault-tolerant statement that metadata is carried out, and utilizes the metasearch Service Database after fault-tolerant, the business information that obtains exporting.
Certainly, fault-tolerant statement can be according to a lot of fault tolerant mechanism settings, such as phonetic, semanteme or both combinations etc., also can be that the degree of correlation according to phonetic and fault-tolerant front and back is to recently determining.The number of fault-tolerant statement is not limit, and can be one, can be for a plurality of yet.
When the number of fault-tolerant statement is one when above, when step 203 is searched for, it is fault-tolerant to select a fault-tolerant statement that metadata is carried out at random, and it is fault-tolerant also can to utilize a plurality of fault-tolerant statements to carry out, and exports each metadata corresponding service data after fault-tolerant.
Fig. 8 is at the fault-tolerant signal of the phonetic of bell musical alphabet.For example, user's Natural Language Search is input as " I want to look for " red " ", and what its real user will look for is " red bean ", and has only " red bean (month circle greet version) " in the business datum.At this moment, fault tolerant mechanism can be fault-tolerant to " red bean " with bell musical alphabet metadata " all red ", and simultaneously, system can generate business datum the True Data " red bean " of " red bean (month circle is greeted version) " with certain filtering policy at the operation initial stage.In ensuing keyword lookup, just can find thing that the user wants.
In the fault-tolerant process of phonetic, also need to cooperate the degree of correlation of fault-tolerant front and back to recently determining fault-tolerant result.As because the polyphone problem, the phonetic of " all red " is HONG_DU, HONG_DOU.If having another first song to cry " coaxing ", its phonetic is HONG_DOU, after present phonetic is fault-tolerant, uses the degree of correlation that other algorithm calculates " red bean " and " all red " and " coaxing ", returns most possible result then.
Having at business datum " singer " in " CRBT search engine " used exists another fault-tolerant, be that synonym is fault-tolerant, referring to shown in Figure 9, this fault-tolerant way refers to that just the user uses other titles such as stage name, place combination name of singer to search for, so the another name information of system by pre-configured part singer head it off well just.
In addition, system also can note " strange language ", safeguards and the optimization reference for template.Can discern most searching request by this scheme, simultaneously remaining Request System be had the prompting of guided bone.
Be that example is further elaborated technical solution of the present invention with the CRBT Natural Language Search below.
Present embodiment is at specific business, realizes the coupling of search natural language by abundant natural language template, thereby locatees wherein keyword; Automatically handle the pinyin marking of keyword by the Chinese pinyin corresponding tables, realize that unisonance is fault-tolerant; By safeguarding that thesaurus realizes synonym coupling, and synon unisonance coupling.
Referring to shown in Figure 4, present embodiment realizes that the detailed process of CRBT Natural Language Search is:
Step 401~402: receive to need the statement of search input, described statement as the keyword lookup Service Database, is judged whether to find, if, process ends then, otherwise, execution in step 403.
Step 403: utilize pre-configured template end to end to this statement is mated filtration, if the match is successful, then execution in step 404, otherwise, execution in step 405.
Step 404: the main body statement that will obtain after will filtering judges whether to find as the keyword lookup Service Database, if, process ends then, otherwise, execution in step 405.
Step 405: utilize pre-configured main body template to the main body statement is mated filtration, if the match is successful, then execution in step 406, otherwise, execution in step 408.
Step 406: judge whether seniority among brothers and sisters, if, then get seniority among brothers and sisters bell sound, process ends, otherwise, execution in step 407.
Step 407: judge whether and need accurately search, if desired, process ends then, otherwise, execution in step 408.
Step 408: need to judge whether full-text search, if desired, process ends then, otherwise, recommend the bell sound.
Referring to shown in Figure 5, a kind of Natural Language Search of the present invention system comprises: natural language filtration unit 51 and natural language R-T unit 52.
Natural language filtration unit 51 is used to preserve metadata corresponding service information, preserves professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template; The regular expression that utilizes natural language template correspondence mates to filter to the statement from natural language R-T unit 52 and obtains metadata, and metadata corresponding service information is exported to natural language R-T unit 52.
Natural language R-T unit 52 is used for the statement of external node equipment is sent to described natural language filtration unit, will export to outside node device from the business datum of natural language filtration unit.
Described natural language filtration unit 51 can comprise: natural language template stores device 511, business datum memory storage 512 and natural language matching control apparatus 513.
Natural language template stores device 511 is used to preserve professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template;
Business datum memory storage 512 is used to preserve metadata corresponding service information;
Natural language matching control apparatus 513, be used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage 511, statement from natural language R-T unit 52 is mated filtration obtain metadata, and the business information that metasearch business datum memory storage 512 obtains is exported to natural language R-T unit 52.
Can comprise in the natural language matching control apparatus 513: natural language matching unit 5131 and fault-tolerant device 5132.
Natural language matching unit 5131, be used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage 511, statement from natural language R-T unit 52 is mated filtration obtain metadata, and metadata is exported to fault-tolerant device 5132.
Fault-tolerant device 5132, be used for searching business datum memory storage 512 and whether contain metadata corresponding service information, if have, directly this business information is exported to natural language R-T unit 52, otherwise, it is fault-tolerant to utilize the fault-tolerant statement of self preserving that metadata is carried out, and utilizes the metasearch business datum memory storage 512 after fault-tolerant to obtain corresponding service information, again this business information is exported to natural language R-T unit 52.
Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims (14)

1, a kind of searching method of natural language is characterized in that, this method may further comprise the steps:
A. be a traffic set natural language modular unit, described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template;
B. at the statement of Natural Language Search input, utilize the regular expression of natural language template correspondence that described statement is mated filtration, obtain metadata;
C. the business information of utilizing the metasearch Service Database to obtain exporting.
2, method according to claim 1, it is characterized in that, each natural language template described in the steps A comprises template and main body template end to end, and described regular expression comprises two parts, and each is template and the main body template part in the corresponding described regular expression respectively end to end.
3, method according to claim 2 is characterized in that, step B comprises:
B11, select a template end to end, utilize this end to end the template correspondence regular expression and end to end template described statement is mated filtration, if coupling is filtered successfully, obtain the body matter of this statement, execution in step B13, otherwise, execution in step B12;
B12, judge end to end template whether all coupling finish, if do not have, execution in step B11 then, otherwise, process ends;
B13, main body template of selection utilize the regular expression of each main body template correspondence, main body template that described body matter is mated filtration, if coupling is filtered successfully, filtered data is a metadata, execution in step C, otherwise, execution in step B14;
B14, judge the main body template whether all coupling finish, if do not have, execution in step B13 then, otherwise, process ends.
4, method according to claim 2 is characterized in that, step B comprises:
B20, with described statement as the keyword lookup Service Database, judge whether to find, if, execution in step C then, otherwise, execution in step B21;
B21, utilize pre-configured template end to end to this statement is mated filtration, if the match is successful, execution in step B22 then, otherwise, with this statement as main body statement, execution in step B23;
B22, the main body statement that obtains after will filtering judge whether to find as the keyword lookup Service Database, if, execution in step C then, otherwise, execution in step B23;
B23, utilize pre-configured main body template that the main body statement is mated filtration, if the match is successful, execution in step C then, otherwise, this statement is accurately searched process ends.
5, method according to claim 3 is characterized in that, among the step B11 end to end template be to select at random in the template of never using end to end or Information Selection according to priority;
The main body template is selection or according to priority Information Selection at random in the main body template of never using among the step B13.
6, method according to claim 5 is characterized in that, described priority is provided with according to character length, the priority height that character is long.
7, method according to claim 1 is characterized in that, this method further comprises: set in advance fault-tolerant statement, then step C comprises:
Judge whether contain metadata corresponding service information in the Service Database, if this business information of direct acquisition is arranged, otherwise it is fault-tolerant to utilize fault-tolerant statement that metadata is carried out, and utilizes the metasearch Service Database after fault-tolerant, the business information that obtains exporting.
8, method according to claim 7 is characterized in that, described fault-tolerant statement is according to phonetic or semantic the setting.
9, method according to claim 7 is characterized in that, described fault-tolerant statement is to recently determining according to the degree of correlation of phonetic and fault-tolerant front and back.
10, method according to claim 7 is characterized in that, the number of described fault-tolerant statement is one or more.
11, method according to claim 10 is characterized in that, when the number of fault-tolerant statement is one when above, is to select a fault-tolerant statement to carry out fault-tolerant to metadata at random among the step C;
When the number of fault-tolerant statement is one when above, is that to utilize more than one fault-tolerant statement to carry out fault-tolerant among the step C, and exports each metadata corresponding service data after fault-tolerant.
12, a kind of search system of natural language is characterized in that, this system comprises:
Natural language filtration unit (51) is used to preserve metadata corresponding service information, preserves professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template; The regular expression that utilizes natural language template correspondence mates to filter to the statement from natural language R-T unit (52) and obtains metadata, and metadata corresponding service information is exported to natural language R-T unit (52);
Natural language R-T unit (52) is used for the statement of external node equipment is sent to natural language filtration unit (51), will export to outside node device from the business information of natural language filtration unit (51).
13, system according to claim 12 is characterized in that, described natural language filtration unit (51) comprising:
Natural language template stores device (511) is used to preserve professional natural language modular unit, and described natural language modular unit comprises an above natural language template, the corresponding regular expression of each natural language template;
Business datum memory storage (512) is used to preserve metadata corresponding service information;
Natural language matching control apparatus (513), be used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage (511), statement from natural language R-T unit (52) is mated filtration obtain metadata, and the business information that metasearch business datum memory storage (512) obtains is exported to natural language R-T unit (52).
14, system according to claim 13 is characterized in that, also comprises in the described natural language matching control apparatus (513):
Natural language matching unit (5131), be used for obtaining the regular expression of the natural language template correspondence of nature language template memory storage (511), statement from natural language R-T unit (52) is mated filtration obtain metadata, and metadata is exported to fault-tolerant device (5132);
Fault-tolerant device (5132), be used for searching business datum memory storage (512) and whether contain metadata corresponding service information, if have, directly this business information is exported to natural language R-T unit (52), otherwise, it is fault-tolerant to utilize the fault-tolerant statement of self preserving that metadata is carried out, and utilizes the metasearch business datum memory storage (512) after fault-tolerant to obtain corresponding service information, this business information is exported to natural language R-T unit (52) again.
CNB2006101040938A 2006-08-01 2006-08-01 Search method and system of a natural language Expired - Fee Related CN100489854C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2006101040938A CN100489854C (en) 2006-08-01 2006-08-01 Search method and system of a natural language

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2006101040938A CN100489854C (en) 2006-08-01 2006-08-01 Search method and system of a natural language

Publications (2)

Publication Number Publication Date
CN1908935A true CN1908935A (en) 2007-02-07
CN100489854C CN100489854C (en) 2009-05-20

Family

ID=37700046

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2006101040938A Expired - Fee Related CN100489854C (en) 2006-08-01 2006-08-01 Search method and system of a natural language

Country Status (1)

Country Link
CN (1) CN100489854C (en)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102402425A (en) * 2010-09-09 2012-04-04 金蝶软件(中国)有限公司 Software concept product generation method and device and ERP generation system
CN101609984B (en) * 2008-06-16 2012-08-29 上海申瑞电力科技股份有限公司 Fast aided modeling method for supervisory control and system
CN101673256B (en) * 2008-09-11 2012-10-31 北大方正集团有限公司 Method and system for automatically extracting article metadata information based on word flow
CN104991943A (en) * 2015-07-10 2015-10-21 百度在线网络技术(北京)有限公司 Music searching method and apparatus
CN105095178A (en) * 2014-05-09 2015-11-25 安徽科大讯飞信息科技股份有限公司 Method and system for realizing text semantics fault-tolerant understanding
CN106599244A (en) * 2016-12-20 2017-04-26 飞狐信息技术(天津)有限公司 Universal original log cleaning device and method
CN107247783A (en) * 2017-06-14 2017-10-13 上海思依暄机器人科技股份有限公司 A kind of method and device of phonetic search music
CN107577755A (en) * 2017-08-31 2018-01-12 江西博瑞彤芸科技有限公司 A kind of searching method
CN108415929A (en) * 2018-01-19 2018-08-17 广州索答信息科技有限公司 A kind of instruction analysis method, electronic equipment and storage medium based on repetition generation technique
CN111897840A (en) * 2020-08-14 2020-11-06 北京字节跳动网络技术有限公司 Data searching method and device, electronic equipment and storage medium
WO2021135103A1 (en) * 2020-05-29 2021-07-08 平安科技(深圳)有限公司 Method and apparatus for semantic analysis, computer device, and storage medium
WO2023165122A1 (en) * 2022-03-04 2023-09-07 康键信息技术(深圳)有限公司 Inquiry template matching method, apparatus and device and storage medium
WO2023231285A1 (en) * 2022-05-30 2023-12-07 青岛海尔科技有限公司 Corpus recognition method and apparatus, and device and storage medium

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101609984B (en) * 2008-06-16 2012-08-29 上海申瑞电力科技股份有限公司 Fast aided modeling method for supervisory control and system
CN101673256B (en) * 2008-09-11 2012-10-31 北大方正集团有限公司 Method and system for automatically extracting article metadata information based on word flow
CN102402425A (en) * 2010-09-09 2012-04-04 金蝶软件(中国)有限公司 Software concept product generation method and device and ERP generation system
CN105095178B (en) * 2014-05-09 2018-08-14 安徽科大讯飞医疗信息技术有限公司 Method and system for realizing text semantic fault-tolerant understanding
CN105095178A (en) * 2014-05-09 2015-11-25 安徽科大讯飞信息科技股份有限公司 Method and system for realizing text semantics fault-tolerant understanding
CN104991943A (en) * 2015-07-10 2015-10-21 百度在线网络技术(北京)有限公司 Music searching method and apparatus
CN106599244A (en) * 2016-12-20 2017-04-26 飞狐信息技术(天津)有限公司 Universal original log cleaning device and method
CN106599244B (en) * 2016-12-20 2024-01-05 飞狐信息技术(天津)有限公司 General original log cleaning device and method
CN107247783A (en) * 2017-06-14 2017-10-13 上海思依暄机器人科技股份有限公司 A kind of method and device of phonetic search music
CN107577755B (en) * 2017-08-31 2020-06-19 江西博瑞彤芸科技有限公司 Searching method
CN107577755A (en) * 2017-08-31 2018-01-12 江西博瑞彤芸科技有限公司 A kind of searching method
CN108415929A (en) * 2018-01-19 2018-08-17 广州索答信息科技有限公司 A kind of instruction analysis method, electronic equipment and storage medium based on repetition generation technique
CN108415929B (en) * 2018-01-19 2021-07-27 广州索答信息科技有限公司 Instruction analysis method based on repeat generation technology, electronic device and storage medium
WO2021135103A1 (en) * 2020-05-29 2021-07-08 平安科技(深圳)有限公司 Method and apparatus for semantic analysis, computer device, and storage medium
CN111897840A (en) * 2020-08-14 2020-11-06 北京字节跳动网络技术有限公司 Data searching method and device, electronic equipment and storage medium
WO2023165122A1 (en) * 2022-03-04 2023-09-07 康键信息技术(深圳)有限公司 Inquiry template matching method, apparatus and device and storage medium
WO2023231285A1 (en) * 2022-05-30 2023-12-07 青岛海尔科技有限公司 Corpus recognition method and apparatus, and device and storage medium

Also Published As

Publication number Publication date
CN100489854C (en) 2009-05-20

Similar Documents

Publication Publication Date Title
CN1908935A (en) Search method and system of a natural language
CN1135485C (en) Identification of words in Japanese text by a computer system
CN101065746A (en) System and method for automatic enrichment of documents
CN1145872C (en) Method for automatically cutting and identiying hand written Chinese characters and system for using said method
CN100339855C (en) Content management system
CN1653448A (en) System and method for searching data sources
CN1912872A (en) Method and system for abstracting new word
CN1617134A (en) System for identifying paraphrases using machine translation techniques
CN1801141A (en) Method and apparatus for translation based on a repository of existing translations
CN1908931A (en) Literal data variable typesetting method
CN101055588A (en) Method for catching limit word information, optimizing output and input method system
CN1896992A (en) Method and device for analyzing XML file based on applied customization
CN1475907A (en) Machine translation system based on examples
CN1094618C (en) Method for self-correction of grammar in machine translation
CN1834955A (en) Multilingual translation memory, translation method, and translation program
CN1164292A (en) Voice-operated services
CN1629833A (en) Method and apparatus for implementing question and answer function and computer-aided write
CN1290901A (en) Method and system for text substitute mode formed by random input source
CN101196898A (en) Method for applying phrase index technology into internet search engine
CN101030267A (en) Automatic question-answering method and system
CN1567303A (en) Method and apparatus for automatic division of structure document information block
CN1661593A (en) Method for translating computer language and translation system
CN110310626A (en) Voice training data creation method, device, equipment and readable storage medium storing program for executing
CN1771494A (en) Automatic segmentation of texts comprising chunsk without separators
CN1410918A (en) Searching engine based on information extraction technique

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20090520

Termination date: 20130801