CN102662953A - Semantic annotation system and method integrated with input method - Google Patents

Semantic annotation system and method integrated with input method Download PDF

Info

Publication number
CN102662953A
CN102662953A CN2012100521438A CN201210052143A CN102662953A CN 102662953 A CN102662953 A CN 102662953A CN 2012100521438 A CN2012100521438 A CN 2012100521438A CN 201210052143 A CN201210052143 A CN 201210052143A CN 102662953 A CN102662953 A CN 102662953A
Authority
CN
China
Prior art keywords
semantic
user
module
editor
storage unit
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
CN2012100521438A
Other languages
Chinese (zh)
Other versions
CN102662953B (en
Inventor
倪旻
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.)
Beijing wisdom Technology Co., Ltd.
Original Assignee
倪旻
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 倪旻 filed Critical 倪旻
Priority to CN201210052143.8A priority Critical patent/CN102662953B/en
Publication of CN102662953A publication Critical patent/CN102662953A/en
Application granted granted Critical
Publication of CN102662953B publication Critical patent/CN102662953B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention discloses a semantic annotation system and a method integrated with input methods. The system comprises: an input method module, an edit idle detection module, a message module, a semantic category identification and management module and a user verification module, wherein semantic analysis of inputted text information which is performed by using the semantic category identification and management module is triggered by the edit idle detection module to extract semantic annotation objects and semantic tags when the edit idle detection module detects that a user is in an edit idle state, and the user is prompted to modify and/or verify the semantic annotation objects and the semantic tags that are extracted automatically by the machine through the user verification module. The system and the method provided in the invention improve acquisition efficiency and accuracy of semantic annotation metadata by integrating input methods with semantic annotation.

Description

With the integrated semantic tagger system and method for input method
Technical field
The present invention relates to computer data and handle and the input field, be specifically related to the semantic tagger system and method integrated with input method.
Background technology
Along with IT technology Web 2.0 technology rapid development particularly; The user produces content and (comprises various types of desktop documents; And a large amount of online document-model blog etc.) quantity increases every day with surprising rapidity; People set forth the viewpoint of oneself, even the complaint suggestion is estimated or expressed to existing product and service by the next idea that exchanges oneself with others of foregoing.
Said user produces content all has high value for individual or commercial undertaking; No matter be that real-time follow-up analysis or later retrieval are looked back, all need technology to retrieve and make things convenient for the location the content of these generations, it is a kind of method that the primitive character in content-based indexes such as keyword that confession retrieves; But let the later stage retrieval need these primitive characters of memory; Memory to the final user will be a challenge, and the people remembers abstract content more easily on the contrary, for instance; Let the people remember that certain concrete name of the dish is difficult to, but remember that the style of cooking is instead easy.Therefore above-mentioned user is produced summary and the mark that content does on the semantic angle and will be very beneficial for later stage searching and locating content.
Document is done semantic tagger from marking the time period of taking place, be divided into and create among the editor and the increase of editor back.Establishment means in the compiling procedure of document and increases semantic label among the editor.It then is after document is accomplished, to increase semantic label through robotization or semi-automatic mode that the editor back increases.For alleviating that the people increases and confirming the workload of semantic label, increase semantic label behind the at present popular normally editor, extract possible label automatically through machine learning, leave the user for and confirm uncertain.No matter adopt which kind of machine learning algorithm; All need the good document of some marks of manual creation as training sample; Therefore to carry out a certain amount of semantic tagger be unavoidable in manual work, and semantic tagger is a dynamic process simultaneously, and error label work for correction amount is also very huge.These work all need be carried out artificial input and semantic tagger.
Existing automatic semanteme marking method is as shown in Figure 1, and this method comprises:
Steps A, obtain new literal paragraph, and be stored in the literal paragraph storage unit;
Step B, this paragraph is carried out grammatical analysis, and store the result into grammatical analysis as a result in the storage unit;
Step C, obtain the semantic tagger plug-in unit according to literal paragraph and the grammatical analysis result of storage and analyze corresponding semantic label, and mark object the most at last and semantic tagger returns.
As a rule, this method realizes through as shown in Figure 2 automatic semantic tagger system.Said system comprises application module, grammer processing module and semantic type of identification administration module, wherein:
Application module is used to obtain new literal paragraph, and is stored in the literal paragraph storage unit;
The grammer processing module is used for this paragraph is carried out grammatical analysis, and stores the result into grammatical analysis as a result in the storage unit;
And a semantic type of identification administration module is used for obtaining the semantic label that the semantic tagger plug-in unit analyzes correspondence according to the literal paragraph of storage with the grammatical analysis result, and marks object the most at last and semantic tagger returns.
Above-mentioned automatic semanteme marking method and system finish the mark of laggard lang justice and return whole section copy editor usually, therefore lack the link that the user confirms usually, and the error that makes automatic semantic tagger occur is difficult to obtain revise, and influences the efficient of semantic tagger.
Therefore how mark is fused among the editor, breaks the whole up into parts, the convenience that the raising system uses and the accuracy rate of semantic tagger are to need the problem of solution at present badly.
Summary of the invention
The objective of the invention is to improve convenience and the accuracy rate that the user carries out semantic tagger.
The invention discloses a kind of and the integrated semantic tagger system of input method, said system comprises:
Input method module is used for carrying out the literal input and the Word message of importing is stored in literal paragraph storage unit;
Edit the idle-detection module, be used to follow the tracks of the information of literal paragraph storage unit, detect the user and whether be in editor's idle condition, and when the user is in editor's idle condition, send editor's idle message to illustrate that the user is in editor's idle condition to message module;
Message module is used for sending the semantic analysis request message according to said editor's idle message to semantic type of identification administration module;
Semantic type of identification administration module; It is right to be used for extracting the mark that comprises preparatory mark object and semantic label in advance based on the Word message that the semantic analysis request message is analyzed said input; Said mark to being saved in preparatory mark object and semantic tagger memory cell, and is confirmed that to the user module sends semantic tagger and confirm request message;
The user confirms module, is used for showing that to the user option of said preparatory mark object and semantic label supplies the user to select based on semantic tagger affirmation request message, and the annotation results after the user is selected to confirm is returned as metadata or additional data storage.
Wherein, Said user confirms that module also comprises semantic tagger modified module and semantic tagger affirmation module; Said semantic tagger modified module is used for showing that to the user Word message of said preparatory mark object and said input supplies the user that preparatory mark object is made amendment that the mark that user's modification is confirmed is stored in the annotation results storage unit;
Said semantic tagger confirms that module is used for showing that to the user option of said semantic label supplies the user to select; Semantic label after the user selected to confirm is stored in the annotation results storage unit, and the annotation results in the annotation results storage unit is returned as metadata or additional data storage.
Wherein, the acquiescence item of the option of the semantic label of confession user selection is the preparatory semantic label that is stored in preparatory mark object and the semantic tagger storage unit.Whether said editor's idle-detection module is in editor's idle condition according to predetermined editor's idle condition judges.
Wherein, Said system also comprises semantic tagger cloud collection module; The annotation results after said semantic tagger cloud collection module is also confirmed user's modification and the contextual information of mark object are uploaded in the extensive mark language material storage unit that stores network side into after weighing in the past.
Wherein, the language material of storing in the said extensive mark language material storage unit is trained for the said semantic type of identification administration module download use of language material model confession at network side or end side.
The invention also discloses a kind of and the integrated semanteme marking method of input method, said method comprises:
Carrying out literal imports and the Word message of importing is stored in the literal paragraph storage unit;
Follow the tracks of the information of literal paragraph storage unit, detect the user and whether be in editor's idle condition, and when the user is in editor's idle condition, be in editor's idle condition with the signal user to message module transmission editor idle message;
Send the semantic analysis request message according to said editor's idle message to semantic type of identification administration module;
It is right that the Word message of analyzing said input based on the semantic analysis request message extracts the mark that comprises preparatory mark object and semantic label in advance; Said mark to being saved in preparatory mark object and semantic tagger memory cell, and is confirmed that to the user module sends semantic tagger and confirm request message;
Show that to the user option of said preparatory mark object and semantic label supplies the user to select based on semantic tagger affirmation request message, the annotation results after the user is selected to confirm is returned as metadata or additional data storage.
Wherein, the user confirms that step further comprises:
Show that to the user Word message of said preparatory mark object and said input supplies the user that preparatory mark object is made amendment, the mark that user's modification is confirmed is stored in the annotation results storage unit;
Show that to the user option of said semantic label supplies the user to select, the semantic label after the user is selected to confirm is stored in the annotation results storage unit, and the annotation results in the annotation results storage unit is returned as metadata or additional data storage.
Wherein, the acquiescence item of the option of the semantic label of confession user selection is the preparatory semantic label that is stored in preparatory mark object and the semantic tagger storage unit.Whether be in editor's idle condition based on predetermined editor's idle condition judges.
Wherein, said method also comprises step: with the contextual information of annotation results after user's modification and the affirmation and mark object, after weighing in the past, upload in the extensive mark language material storage unit that stores network side into.
Wherein, said method also comprises: the language material of storing in the said extensive mark language material storage unit is trained for the said semantic type of identification administration module download use of language material model confession at network side or end side.
The present invention is through becoming one semantic tagger and input method; Realized that the prompting user carries out the manual work affirmation for automatic semantic analysis result in user's input characters process; Improved metadata greatly and obtained efficient and accuracy rate, simultaneously, increased mark object modification and network and shared and collaboration feature; Expand the range of application of semantic tagger system, further improved the availability of system.
Description of drawings
Fig. 1 is the method flow diagram of existing automatic semanteme marking method;
Fig. 2 is the block diagram of existing automatic semantic tagger system;
Fig. 3 is first embodiment of the invention and the block diagram integrated semantic tagger system of input method;
Fig. 4 is first embodiment of the invention and the process flow diagram integrated semanteme marking method of input method;
Fig. 5 is the synoptic diagram that the user of first embodiment of the invention confirms the interface;
Fig. 6 is second embodiment of the invention and the block diagram integrated semantic tagger system of input method;
Fig. 7 is the synoptic diagram that the user of second embodiment of the invention confirms the interface;
Fig. 8 is third embodiment of the invention and the block diagram integrated semantic tagger system of input method.
Embodiment
Further specify embodiment of the present invention below in conjunction with accompanying drawing.
The input of Any Application all can such as for Chinese text, just need Chinese character coding input method by means of specific input media, and the information of all inputs all can be input to corresponding application program such as the Word of Microsoft process software through input method.Present embodiment with semantic tagger and input method carry out integrated in, so just accomplished the mark process is jumped out application program, and will mark and use decoupling zero.The preparatory markup information that obtains by means of Machine Method is presented to before the application system, just as selecting the input speech, lets the user confirm.To realize that the user confirms link.Make thus when the user uses semanteme is explained process and the literal input combines together, therefore can improve mark efficient and accuracy.
Fig. 3 is first embodiment of the invention and the block diagram integrated semantic tagger system of input method.Said system comprises that input method module, user confirm module, editor's idle-detection module, message processing module and semantic type of identification administration module.
Wherein, input method module is used to carry out the literal input, and the Word message of input is stored in the literal paragraph storage unit.
Editor's idle-detection module is moved the information of following the tracks of literal paragraph storage unit always; Detect the user and whether be in editor's idle condition; If the user is in editor's idle condition; For example the user at the fixed time in the section not input characters or user input show sentence or paragraph carried out the punctuation mark (like comma, fullstop, branch etc.) that sense-group is cut apart, then edit the idle-detection module and send editor's idle message to message module and be in editor's idle condition with the signal user.
Message module is sent the semantic analysis request message according to said editor's idle message to semantic type of identification administration module.
Semantic type of identification administration module according to semantic analysis request message analyzing stored in literal paragraph storage unit the literal paragraph and extract all semantic taggers to (comprising semantic tagger object and semantic label); And be saved in preparatory mark object and semantic tagger storage unit, and confirm module transmission semantic tagger affirmation request message to the user.
The user confirms that module comprises semantic tagger affirmation module; Semantic tagger confirms that module can start the user according to semantic tagger affirmation request message and confirm processing procedure; Show that to the user option of mark object and semantic label supplies the user to select; Default option wherein is stored in the preparatory mark label in preparatory mark object and the semantic tagger storage unit for semantic type of identification administration module; Information after the user selects to confirm is saved in the annotation results storage unit, and returns annotation results as metadata or additional data storage.
Preferably, the mode of message transmission can adopt editor's idle message storage unit, the semantic analysis request message storage unit that is used to store the semantic analysis request message that is provided for storing editor's idle message, the semantic tagger affirmation request message storage unit that is used to store semantic tagger affirmation request message.Each processing module is sent message and is obtained message through the information content that changes above-mentioned message storage and the variation of monitoring the information content of corresponding message storage.For example; Editor's idle-detection module detects the User Status of upgrading in said editor's idle message storage unit when the user is in editor's idle condition and is the free time; Message module detects said editor's idle message storage unit User Status and changes into the free time, then upgrades the information in the semantic analysis request message storage unit.
Fig. 4 is first embodiment of the invention and the process flow diagram integrated semanteme marking method of input method.Said method comprises:
Step 100, carry out literal input, the Word message of input is stored in the literal paragraph storage unit;
Step 200, follow the tracks of the information of literal paragraph storage unit, and detect the user and whether be in editor's idle condition,, then send editor's idle message and be in editor's idle condition with the signal user to message module if the user is in editor's idle condition;
Step 300, based on said editor's idle message, send the semantic analysis request message to a semantic type of identification administration module;
Step 400, based on semantic analysis request message analyzing stored in literal paragraph memory cell the literal paragraph and extract all semantic taggers to (comprising semantic tagger object and semantic label); And be saved in preparatory mark object and semantic tagger memory cell, and confirm module transmission semantic tagger affirmation request message to semantic tagger;
Step 500, go people's request message to start the user according to semantic tagger to confirm processing procedure, the information after the user is confirmed is saved in the annotation results storage unit and with annotation results and is stored in metadata or the additional data.
The flow process of the semanteme marking method of first embodiment of the invention below is described by way of example.
(1) user uses document editor (for example, the Word of Microsoft editing machine) Edit Document, uses input method to import following content:
L=" I am from Henan, "
(2) editor's idle-detection module can judge whether to be in editor's idle condition according to specific policy; For example in the time of user's inputting punctuation mark; It is idle to be editor, and so according to the input in the last step, said editor's idle-detection module can trigger editor's idle message.
M Idle={ editor is idle, Word}
(3) message module can produce a semantic analysis request message after trapping this editor's idle message, triggers semantic type of identification administration module accordingly.
(4) semantic type of identification administration module analyzed the information among the L, obtains following semantic analysis result:
R Can={ " place ": " Henan ", start=3, length=2}
(5) semantic analysis result will be delivered to the user and confirm module, confirm by the user that module generates and ejects and as shown in Figure 5 select the similar user's acknowledgement window of vocabulary with input method, and this window is divided into two parts; Top shows mark object " Henan ", and the bottom shows the option that semantic tagger is corresponding, and such as " name; place name; mechanism's name or the like ", what acquiescence was chosen is type corresponding among the R, and it is preferred that this example corresponds to " place name ".
After the user confirmed, the user confirmed that module can produce following annotation results and also this result returned to storage unit:
R={ " place ": " Henan " }
Finally by the user confirm module obtain said annotation results and be stored into metadata corresponding or additional data in.
In semantic analysis process, classification identification inevitably mistake can occur automatically.Semantic analysis classification identification error is divided into two types usually, and first semantic type profiling error is such as the place name that is identified as that should be name; Other one type is semantic tagger object identification error, such as being triliteral two words that are identified as.For Error type I, confirm module through the semantic tagger of introducing in the first embodiment of the invention, can repair.And for error type II, propose the second embodiment of the present invention at this this is done further to optimize.
The block diagram with the integrated semantic tagger system of input method of second embodiment of the invention is as shown in Figure 6.Said system confirms to have increased in the module semantic tagger modified module the user on the basis of first embodiment.In a second embodiment; The user confirms that module comprises semantic tagger affirmation module and semantic tagger modified module; Said semantic tagger modified module confirms that according to semantic tagger request message obtains the literal paragraph of input and marks object in advance; The literal paragraph of input is played up with preparatory mark object and semantic tagger information mixed; Show the literal paragraph of input and mark object in advance simultaneously, confirm to mark object, and the mark object storage that will confirm according to the result of user's modification or affirmation is to the annotation results storage unit by the user.Said semantic tagger confirms that module is used to show that the option of semantic label supplies the user to select; Default option wherein is stored in the preparatory mark label in preparatory mark object and the semantic tagger storage unit for semantic type of identification administration module, and the information after the user selects to confirm is saved in the annotation results storage unit returns annotation results as metadata store.
The flow process of the semanteme marking method of second embodiment of the invention below is described by way of example.
(1) user comes Edit Document using document editor (for example, the Word of Microsoft editing machine), uses input method to import following content:
L=" I have visited Palace Museum today, "
(2) editor's idle-detection module can judge whether to be in editor's idle condition according to specific policy; Such as when user's inputting punctuation mark the time; It is idle to be editor, and so according to the input in the last step, said editor's idle-detection module can trigger editor's idle message.
M Idle={ editor is idle, Word}
(3) message module can produce a semantic analysis request message after trapping this editor's idle message, triggers semantic type of identification administration module accordingly.
(4) semantic type of identification administration module can be analyzed the information among the L, can obtain following semantic analysis result:
R Can={ " mechanism's name ": " the Forbidden City ", start=5, length=2}
(5) user confirms that module obtains the Word message of semantic analysis result and user input, ejects user's acknowledgement window as shown in Figure 7, and this window is divided into two parts; Top show in advance mark object " the Forbidden City " with and contextual information, wherein mark object in advance through the bright demonstration of high transom window height, the reference position of this high transom window is all adjustable; The user can revise the mark object through revising this high transom window, and the bottom shows the option that semantic tagger is corresponding, and such as " name; place name; mechanism's name or the like ", what acquiescence was chosen is the preparatory semantic label in the semantic analysis result, and it is preferred that this example corresponds to " mechanism's name ".
Object should be " Palace Museum " if the user thinks mark, promptly can museum be added through revising this high transom position of window, and after the user confirmed, the user confirmed that module will produce following annotation results and also this result returned to storage unit:
R={ " mechanism's name ": " Palace Museum " }
(6) the semantic tagger result who confirms through the user is confirmed by the user that finally module obtains and be stored in metadata corresponding or the additional data.
Guarantee to improve artificial confirm efficient in, be that system increases the network sharing characteristic, make that can multiple-person cooperative work sharing the result also is one of demand for semantic tagger of the present invention system.Propose the third embodiment of the present invention at this present invention is done further optimization.
Fig. 8 is the 3rd embodiment and the block diagram integrated semantic tagger system of input method.Said system has increased semantic tagger cloud collection module on the basis of second embodiment; Mark language material after semantic tagger cloud collection module is also confirmed user's modification; Comprise the contextual information that marks object, after weighing in the past, upload in the extensive mark language material storage unit that stores network side into.Being stored in language material in the extensive mark language material storage unit becomes the semantic analysis model at network side through model training device direct Training the language material model of semantic type of identification administration module is carried out model modification, thereby perhaps language material is distributed to the renewal that semantic type of identification administration module is trained to language material by the model training device that embeds in this module language material model implementation model.
The present invention is through becoming one semantic tagger and input method; Realized that the prompting user carries out the manual work affirmation for automatic semantic analysis result in user's input characters process, improved user's metadata greatly and obtained efficient, simultaneously; Having increased mark object modification and network shares and collaboration feature; Expand the range of application of semantic tagger system, realized obtaining the possibility of high-quality metadata language material in enormous quantities, further improved the availability of system.
Above-mentioned preferred embodiment of the present invention and the institute's application technology principle of being merely, any technician who is familiar with the present technique field is in the technical scope that the present invention discloses, and the variation that can expect easily or replacement all should be encompassed in protection scope of the present invention.

Claims (10)

  1. One kind with the integrated semantic tagger system of input method, said system comprises:
    Input method module is used for carrying out the literal input and the Word message of importing is stored in literal paragraph storage unit;
    Edit the idle-detection module, be used to follow the tracks of the information of literal paragraph storage unit, detect the user and whether be in editor's idle condition, and when the user is in editor's idle condition, send editor's idle message to illustrate that the user is in editor's idle condition to message module;
    Message module is used for sending the semantic analysis request message according to said editor's idle message to semantic type of identification administration module;
    Semantic type of identification administration module; It is right to be used for extracting the mark that comprises preparatory mark object and semantic label in advance based on the Word message that the semantic analysis request message is analyzed said input; Said mark to being saved in preparatory mark object and semantic tagger memory cell, and is confirmed that to the user module sends semantic tagger and confirm request message;
    The user confirms module, is used for showing that to the user option of said preparatory mark object and semantic label supplies the user to select based on semantic tagger affirmation request message, and the annotation results after the user is selected to confirm is returned as metadata or additional data storage.
  2. 2. as claimed in claim 1 and the integrated semantic tagger system of input method; It is characterized in that; Said user confirms that module also comprises semantic tagger modified module and semantic tagger affirmation module; Said semantic tagger modified module is used for showing that to the user Word message of said preparatory mark object and said input supplies the user that preparatory mark object is made amendment that the mark that user's modification is confirmed is stored in the annotation results storage unit;
    Said semantic tagger confirms that module is used for showing that to the user option of said semantic label supplies the user to select; Semantic label after the user selected to confirm is stored in the annotation results storage unit, and the annotation results in the annotation results storage unit is returned as metadata or additional data storage.
  3. According to claim 1 or claim 2 with the integrated semantic tagger system of input method, it is characterized in that supplying the acquiescence item of the option of the semantic label that the user selects is the preparatory semantic label that is stored in preparatory mark object and the semantic tagger storage unit;
    Whether said editor's idle-detection module is in editor's idle condition according to predetermined editor's idle condition judges.
  4. According to claim 1 or claim 2 with the integrated semantic tagger system of input method; It is characterized in that; Said system also comprises semantic tagger cloud collection module; The annotation results after said semantic tagger cloud collection module is also confirmed user's modification and the contextual information of mark object are uploaded in the extensive mark language material storage unit that stores network side into after weighing in the past.
  5. 5. as claimed in claim 4 and the integrated semantic tagger system of input method is characterized in that, the language material of storing in the said extensive mark language material storage unit is trained for the language material model in network side or end side and supplies a said semantic type of identification administration module to download to use.
  6. One kind with the integrated semanteme marking method of input method, said method comprises:
    Carrying out literal imports and the Word message of importing is stored in the literal paragraph storage unit;
    Follow the tracks of the information of literal paragraph storage unit, detect the user and whether be in editor's idle condition, and transmission editor idle message is in editor's idle condition with the signal user when the user is in editor's idle condition;
    Send the semantic analysis request message based on said editor's idle message;
    It is right that the Word message of analyzing said input based on the semantic analysis request message extracts the mark that comprises preparatory mark object and semantic label in advance, and said mark to being saved in preparatory mark object and semantic tagger memory cell, and is sent semantic tagger affirmation request message;
    Show that to the user option of said preparatory mark object and semantic label supplies the user to select based on semantic tagger affirmation request message, the annotation results after the user is selected to confirm is returned as metadata or additional data storage.
  7. 7. as claimed in claim 6 and the integrated semanteme marking method of input method is characterized in that the user confirms that step further comprises:
    Show that to the user Word message of said preparatory mark object and said input supplies the user that preparatory mark object is made amendment, the annotation results that user's modification is confirmed is stored in the annotation results storage unit;
    Show that to the user option of said semantic label supplies the user to select, the semantic label after the user is selected to confirm is stored in the annotation results storage unit, and the annotation results in the annotation results storage unit is returned as metadata or additional data storage.
  8. 8. like claim 6 or the 7 described and integrated semanteme marking methods of input method, it is characterized in that the acquiescence item of the option of the semantic label that the confession user selects is the preparatory semantic label that is stored in preparatory mark object and the semantic tagger storage unit; And
    Whether be in editor's idle condition based on predetermined editor's idle condition judges.
  9. 9. like claim 6 or the 7 described and integrated semanteme marking methods of input method, it is characterized in that said method also comprises step:
    With the contextual information of annotation results after user's modification and the affirmation and mark object, after the past is heavy, upload in the extensive mark language material memory cell that stores network side into.
  10. 10. as claimed in claim 9 and the integrated semanteme marking method of input method is characterized in that, also comprise:
    The language material of storing in the said extensive mark language material storage unit is trained for the said semantic type of identification administration module download use of language material model confession at network side or end side.
CN201210052143.8A 2012-03-01 2012-03-01 With the semantic tagger system and method that input method is integrated Active CN102662953B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210052143.8A CN102662953B (en) 2012-03-01 2012-03-01 With the semantic tagger system and method that input method is integrated

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210052143.8A CN102662953B (en) 2012-03-01 2012-03-01 With the semantic tagger system and method that input method is integrated

Publications (2)

Publication Number Publication Date
CN102662953A true CN102662953A (en) 2012-09-12
CN102662953B CN102662953B (en) 2016-04-06

Family

ID=46772444

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210052143.8A Active CN102662953B (en) 2012-03-01 2012-03-01 With the semantic tagger system and method that input method is integrated

Country Status (1)

Country Link
CN (1) CN102662953B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105068999A (en) * 2015-08-14 2015-11-18 浪潮集团有限公司 Method and apparatus for identifying amended entity words
CN109683773A (en) * 2017-10-19 2019-04-26 北京国双科技有限公司 Corpus labeling method and device
CN109753976A (en) * 2017-11-01 2019-05-14 中国电信股份有限公司 Corpus labeling device and method
CN110807486A (en) * 2019-10-31 2020-02-18 北京达佳互联信息技术有限公司 Method and device for generating category label, electronic equipment and storage medium
CN113127635A (en) * 2019-12-31 2021-07-16 阿里巴巴集团控股有限公司 Data processing method, device and system, storage medium and electronic equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040267774A1 (en) * 2003-06-30 2004-12-30 Ibm Corporation Multi-modal fusion in content-based retrieval
CN1936892A (en) * 2006-10-17 2007-03-28 浙江大学 Image content semanteme marking method
CN101075230A (en) * 2006-05-18 2007-11-21 中国科学院自动化研究所 Method and device for translating Chinese organization name based on word block
CN101216819A (en) * 2007-12-28 2008-07-09 北京邮电大学 Name card information Chinese to English automatic translation method based on domain ontology
CN101334796A (en) * 2008-02-29 2008-12-31 浙江师范大学 Personalized and synergistic integration network multimedia search and enquiry method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040267774A1 (en) * 2003-06-30 2004-12-30 Ibm Corporation Multi-modal fusion in content-based retrieval
CN101075230A (en) * 2006-05-18 2007-11-21 中国科学院自动化研究所 Method and device for translating Chinese organization name based on word block
CN1936892A (en) * 2006-10-17 2007-03-28 浙江大学 Image content semanteme marking method
CN101216819A (en) * 2007-12-28 2008-07-09 北京邮电大学 Name card information Chinese to English automatic translation method based on domain ontology
CN101334796A (en) * 2008-02-29 2008-12-31 浙江师范大学 Personalized and synergistic integration network multimedia search and enquiry method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘挺等: "基于最大熵分类器的语义角色标注", 《软件学报》 *
刘长松等: "用统计方法实现汉字输入的智能联想", 《中文信息学报》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105068999A (en) * 2015-08-14 2015-11-18 浪潮集团有限公司 Method and apparatus for identifying amended entity words
CN109683773A (en) * 2017-10-19 2019-04-26 北京国双科技有限公司 Corpus labeling method and device
CN109753976A (en) * 2017-11-01 2019-05-14 中国电信股份有限公司 Corpus labeling device and method
CN109753976B (en) * 2017-11-01 2021-03-19 中国电信股份有限公司 Corpus labeling device and method
CN110807486A (en) * 2019-10-31 2020-02-18 北京达佳互联信息技术有限公司 Method and device for generating category label, electronic equipment and storage medium
CN110807486B (en) * 2019-10-31 2022-09-02 北京达佳互联信息技术有限公司 Method and device for generating category label, electronic equipment and storage medium
CN113127635A (en) * 2019-12-31 2021-07-16 阿里巴巴集团控股有限公司 Data processing method, device and system, storage medium and electronic equipment
CN113127635B (en) * 2019-12-31 2024-04-02 阿里巴巴集团控股有限公司 Data processing method, device and system, storage medium and electronic equipment

Also Published As

Publication number Publication date
CN102662953B (en) 2016-04-06

Similar Documents

Publication Publication Date Title
US10872104B2 (en) Method and apparatus for natural language query in a workspace analytics system
WO2018032937A1 (en) Method and apparatus for classifying text information
CN107392143A (en) A kind of resume accurate Analysis method based on SVM text classifications
US10891430B2 (en) Semi-automated methods for translating structured document content to chat-based interaction
CN102880649A (en) Individualized information processing method and system
CN102654874A (en) Bill data management method and system
CN101490668A (en) Reuse of available source data and localizations
CN102662953B (en) With the semantic tagger system and method that input method is integrated
CN104199871A (en) High-speed test question inputting method for intelligent teaching
CN102591475A (en) Content input method and system for online editor
CN105138575A (en) Analysis method and device of voice text string
CN104169912A (en) Information processing terminal and method, and information management apparatus and method
CN103559234A (en) System and method for automated semantic annotation of RESTful Web services
CN105893626A (en) Index library creation method used for nuclear power engineering and index system adopting index library creation method
CN104484379B (en) Determine the method and apparatus of music property relationship and inquiry processing method and device
CN101980184A (en) Method and system for annotating contents of documents
US20140075299A1 (en) Systems and methods for generating extraction models
CN104731874A (en) Evaluation information generation method and device
CN101452443A (en) Recording medium for recording logical structure model creation assistance program, logical structure model creation assistance device and logical structure model creation assistance method
CN110413882B (en) Information pushing method, device and equipment
JP2006268690A (en) Faq presentation/improvement method, faq presentation/improvement device and faq presentation/improvement program
CN111191429A (en) System and method for automatic filling of data table
Benajiba et al. A web application for dialectal Arabic text annotation
CN111753536A (en) Automatic patent application text writing method and device
CN110413307A (en) Correlating method, device and the electronic equipment of code function

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
C41 Transfer of patent application or patent right or utility model
TR01 Transfer of patent right

Effective date of registration: 20161227

Address after: 100085 Beijing, Haidian District, No. ten on the ground floor, No. 1, No. 4 international building, No. 2206

Patentee after: Beijing wisdom Technology Co., Ltd.

Address before: 100094 Haidian District North Road, Beijing, No. 68

Patentee before: Ni Min