CN101719122A - Method for extracting Chinese named entity from text data - Google Patents
Method for extracting Chinese named entity from text data Download PDFInfo
- Publication number
- CN101719122A CN101719122A CN200910227302A CN200910227302A CN101719122A CN 101719122 A CN101719122 A CN 101719122A CN 200910227302 A CN200910227302 A CN 200910227302A CN 200910227302 A CN200910227302 A CN 200910227302A CN 101719122 A CN101719122 A CN 101719122A
- Authority
- CN
- China
- Prior art keywords
- rule
- named entity
- entity
- chinese
- name
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Machine Translation (AREA)
Abstract
The invention discloses a method for extracting a Chinese named entity from text data. The method comprises the following steps of: dividing Chinese words; removing Chinese stopwords; analyzing error types and reasons of the named entities in the Chinese word segmentation result; establishing modification rules respectively according to the error types and reasons; according to the labeled standard corpus matching result, outputting the correct named entities, further modifying the rules for the named entities where errors occur, inputting the modified rules to a rule set, and updating the rule set; and according to the labeled standards, continuously modifying the rules until the output result is optimum, and determining the optimum rule set. The method improves the accuracy rate of extracting the named entities and ensures extraction efficiency; and the method is applied in fields of network information treatment, network data mining, information security and the like, and can provide a good pretreatment foundation for various subsequent treatments.
Description
Technical field
The present invention relates to network information extraction and field of information processing, especially relate to a kind of method of from text data, extracting Chinese named entity.
Background technology
Along with popularizing of network, web page text has carried the most network information as a kind of important information carrier.Named entity described herein refers to phrases such as name in the web page text, place name, organizational structure's name, time, and these phrases all are information elements basic in the text, have often indicated the main contents of article, are the bases of correct understanding text.Therefore, effective extraction of named entity has very important significance for efficiently obtaining info web.For example, if in information extraction, do not extract entity earlier, not may discern entity relationship, also impossible extraction incident masterplate; In digest generates, many times be to fixed mode filling, fill content and mostly be " who ", " when ", " where " or the like, this is the content of named entity just, therefore obtains the extraction that these contents just be unable to do without named entity from article.Therefore, the accurate extraction of named entity is the prerequisite of text understanding, is the basis of all follow-up works of text information processing field.Yet the difficulty of named entity extraction work is: at real text Chinese sentence is not unit with the speech, but is unit with the word.In order to reduce the complexity that Chinese named entity is extracted, usually a minute word information is used for Chinese named entity and extracts, but, can lead to errors and spread if the mistake of participle can't obtain correcting in the named entity leaching process.The mistake that named entity extracts mainly is divided into two classes: Error type I is the misjudgment of named entity border.A kind of situation of this mistake is to have lost the part that belongs to named entity originally, when generally occurring in the place name of extracting the long or structure more complicated of length and mechanism's name.For example: " airport, Tananarive " is extracted as " Tananarive ", lost " airport " this suffix.Another kind of situation is exactly that this word or speech that does not belong to this named entity has been comprised; For example: the place name " PORT OF WANXUAN " in " PORT OF WANXUAN organizes more than 30 ship to drop into rescue work " the words is thought mechanism's name " PORT OF WANXUAN tissue ".Error type II is the misjudgment of named entity type, and for example: it is name that place name " Liu Zhuan " mistake is known.The generation of this mistake generally all is because two kinds of named entities may have similar place on feature.Also be one of surname of name in the place name " Liu Zhuan " in the example as above, so caused this mistake.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of method of from text data, extracting Chinese named entity, improve named entity and extract accuracy rate, and can guarantee extraction efficiency.
For achieving the above object, the present invention by the following technical solutions:
The method of extracting Chinese named entity from text data of the present invention may further comprise the steps:
A. Chinese word segmentation;
B. Chinese stop words is removed;
C. analyze named entity type of error and reason among the Chinese word segmentation result;
D. formulate modification rule respectively at type of error and reason;
E. according to labeled standard corpus matching result,, named entity that mistake occurs further to the rule correction, and is input to rule set, the update rule collection with revised rule with correct named entity output;
F. constantly carry out rule correction according to labeled standards, up to the output result optimal, and definite optimum rule set.
Further, the modification rule described in the d step comprises and merges rule, finger name extracting rule, border modification rule and type modification rule together.
Further, described merging rule is about to should belong to an entity together among the Chinese word segmentation result and is that an entity merged in two or more words by false segmentation; The described name extracting rule that refers to together promptly finds the speech that refers to same name in the text and unifies mark; Described border modification rule is promptly revised the mistake of having lost self part when named entity extracts, if place name is lost suffix, then sets up corresponding place name suffix dictionary update information is provided; Described type modification rule, the type misjudgment when promptly revising the named entity extraction.
The invention has the beneficial effects as follows:
Method of the present invention is on the basis of Chinese word segmentation, the type of error that occurs when extracting at named entity and and reason, the Chinese named entity extracting method of a kind of practicability of proposition; This method proposes to set up some rules according to analyzing the rule that Chinese named entity occurs, and the mistake that these rules occur in can extracting named entity is targetedly effectively revised; By checking each bar rule is optimized screening then, forms the optimal rules storehouse, the mistake that occurs during Chinese named entity is extracted is revised, and finally realizes effective extraction of Chinese named entity; The present invention casts aside some loaded down with trivial details algorithms in named entity extracts, the analysis that adds in rule-based named entity extracting method rule reaches the modification rule of formulating therefrom, revise the mistake among the Chinese word segmentation result, in real time modification rule is constantly adjusted according to labeled standards simultaneously, make modification rule reach optimum, the named entity result who is extracted accurately and efficiently; Compare with traditional named entity extracting method, the characteristics of the inventive method are: 1, carry out on the basis of Chinese word segmentation, guaranteed that institute's analytic target is speech rather than single word, this is in the high efficiency that has guaranteed entity extraction basically; 2, the entity extraction type of error is carried out labor, formulated modification rule targetedly and adjust the extraction mistake, guaranteed the accuracy of entity extraction like this.
Other advantages of the present invention, target and feature will be set forth to a certain extent in the following description, and to a certain extent, based on being conspicuous to those skilled in the art, perhaps can obtain instruction from the practice of the present invention to investigating hereinafter; Target of the present invention and other advantages can realize and obtain by specifically noted mode in following instructions and the accompanying drawing.
Description of drawings
Accompanying drawing is the operational flowchart of the inventive method.
Embodiment
Below in conjunction with drawings and Examples the present invention is further described.
New method of the present invention is at first carried out labor to the named entity word-building rule, then on basis to Chinese word segmentation and stop words removal, form structure and context environmental according to entity instance, formulate matched rule and revise the named entity extraction mistake that the Chinese word segmentation mistake causes, the result will be adjusted at last and labeled standards is compared, the further regulation rule of wrong entity is revised, continuous like this rule is screened and upgraded, form the optimal rules storehouse at last named entity is extracted.In conjunction with the accompanying drawings, the method for extracting Chinese named entity from text data of the present invention may further comprise the steps:
A. Chinese word segmentation;
B. Chinese stop words is removed;
C. analyze named entity type of error and reason among the Chinese word segmentation result;
D. formulate four kinds of modification rules respectively at type of error and reason, promptly merge rule, finger name extracting rule, border modification rule and type modification rule together; Described merging rule is about to should belong to an entity together among the Chinese word segmentation result and is that an entity merged in two or more words by false segmentation; The described name extracting rule that refers to together promptly finds the speech that refers to same name in the text and unifies mark; Described border modification rule is promptly revised the mistake of having lost self part when named entity extracts, if place name is lost suffix, then sets up corresponding place name suffix dictionary update information is provided; Described type modification rule, the type misjudgment when promptly revising the named entity extraction.
E. according to labeled standard corpus matching result,, named entity that mistake occurs further to the rule correction, and is input to rule set, the update rule collection with revised rule with correct named entity output;
F. constantly carry out rule correction according to labeled standards, up to the output result optimal, and definite optimum rule set.
Below be described in further detail the particular content that rule is revised:
Because the form of date and time is more fixing, can extract by setting up finte-state machine, so the rule that this method is set up mainly is to be used to revise some mistakes that take place when Chinese name, place name and mechanism's name are extracted more accurately.The rule that wherein is used to extract named entity is broadly divided into four classes, is described respectively below and illustrates.
Rule classification 1: merge rule
This rule is devoted to revise two kinds of mistakes in the named entity extraction:
Error of the first kind: the short-lived name entity that the long life name entity that the handle that exists in the named entity leaching process belongs to an integral body is divided into several successive extracts.For example: " China Radio International " (mechanism's name) is identified as " China " (place name) and two named entities of " international broadcast station " (mechanism's name).
Second kind of mistake: two continuous named entities that belong to dominance relation do not merge.According in the named entity recognition standard at national basis resource evaluation and test center about the predominate structure rule, be the named entity of dominance relation if the structure that continues then is labeled as one when two.For example: " TaiWan, China " should be labeled as a named entity, and should not be labeled as two named entities in " China " " Taiwan ".
Rule classification 2: refer to the name extracting rule together
This rule is intended to find the speech that refers to same name, and unified mark.
It is relatively easy that the same finger of the foreign name of Chinese is judged.The full name of general foreign name has " " interval as first name and last name, for example: " bielke Islington ".So when identification refers to name together, generally all be name or the surname of seeking the name full name.For example: " sherry " and " sherry De Leipa ", " Jordon " and " Michael Jordon " in same section literal refers to a people together.
Chinese personal name is because composition form is relatively various, so rule is set also more complicated.
Chinese personal name refers to judgment rule together: (1) entity 2 is name parts of entity 1, and for example: entity 1 is " Liu Yunfei ", and entity 2 is " cloud flies "; (2) entity 2 is surname parts of entity 1, and for example: entity 1 is " Li Rongbiao ", and entity 2 is " Lee "; (3) entity 2 is that the surname of entity 1 partly adds the name suffix, and for example: entity 1 is " Zhang Tianjiao ", and entity 2 is " opening total "; (4) entity 2 is that the name part of entity 1 adds the name suffix, and for example: entity 1 is " Liu Dehua ", and entity 2 is " China is young "; (5) entity 2 is surname parts that the name prefix adds entity 1, and for example: entity 1 is " Li Zeming ", and entity 2 is " Lao Li "; (6) entity 2 is name parts that the name prefix adds entity 1, and for example: entity 1 is " Chen Ye ", and entity 2 is " little firelight or sunlight "; (7) entity 2 is that the surname of entity 1 partly adds appellation, and for example: entity 1 is " Mao Zedong ", and entity 2 is " Chairman Mao ".
Rule classification 3: border modification rule
This rule is mainly used in revises the mistake of having lost self part when named entity extracts.This type of wrong most of situation that takes place all is that place name is lost suffix.At this situation, we can set up a place name suffix dictionary update information is provided.For example: lost suffix " town " in the time of identification " stupe Li Zhen ", it is " stupe profit " that mistake is known.
Rule classification 4: type modification rule
Type misjudgment when this rule is used to revise the named entity extraction, for example: place name " Benin " is identified as name.The Else Rule of comparing, this regular correction effect is less better, because the relatively fuzzyyer named entity feature of type is not clearly, utilizes the method for rule to be not easy to be applicable to all situations, tends to take place mistake correction.
Every rule in the rule base is formed us and is used for reference the form of SEGTAG system, specifically suc as formula mistake! Do not find Reference source.：
POST_LIST+CONTEXT_WORD →＜TYPE〉RESULTNE＜/TYPE 〉, the formula mistake! Do not find Reference source.In, POST_LIST is meant the match pattern that is made of jointly part of speech sequence, inner keyword and suffix feature speech etc., wherein inner keyword and suffix feature speech are optional in a rule; CONTEXT_WORD refers to the deictic words before and after the named entity, also is an option; RESULTNE refers to the named entity that finally identifies after the correction, and TYPE refers to the type of the named entity that finally identifies.Provide giving an example of some rules below:
＜1〉*/nr1#/nr2 →＜PER * #＜/PER, for example: " Liu/nr1 moral China/nr2 " →＜PER Liu Dehua＜/PER;
＜2〉*/ns#/n university/n →＜ORG * # university＜/ORG, for example: " Beijing/ns post and telecommunications/n university/n " →＜ORG Beijing University of Post ﹠ Telecommunication＜/ORG;
＜3〉*/ns#/ns be/v →＜LOC * #＜/LOC, for example: " China/ns Hong Kong/ns is/bright Oriental Pearl of v." → "＜LOC〉Hong-Kong＜/LOC〉be a bright Oriental Pearl.”
In addition, for every in rule base rule has all been set priority, in coupling,, then be as the criterion with that high rule of priority if a named entity has mated many rules.
Make up rule base according to the rule of being formulated and come the entity that is extracted is revised, and correction result and labeled standards are compared, rule base is continued to optimize adjustment, finally form the extraction of optimal rules storehouse realization named entity according to correction result.
Use new method of the present invention to extract named entity, can be from avoiding the adverse effect that individual character brings entity extraction, improve the accuracy and the high efficiency of entity extraction greatly, it extracts the result and also provides strong assurance for the excavation of various processing of the text in later stage and web page text data.The present invention is applicable to fields such as network information processing, network data excavation and information security, and good pre-service basis can be provided for the various processing in later stage.
Explanation is at last, above embodiment is only unrestricted in order to technical scheme of the present invention to be described, other modifications that those of ordinary skills make technical scheme of the present invention or be equal to replacement, only otherwise break away from the spirit and scope of technical solution of the present invention, all should be encompassed in the middle of the claim scope of the present invention.
Claims (3)
1. method of extracting Chinese named entity from text data is characterized in that this method may further comprise the steps:
A. Chinese word segmentation;
B. Chinese stop words is removed;
C. analyze named entity type of error and reason among the Chinese word segmentation result;
D. formulate modification rule respectively at type of error and reason;
E. according to labeled standard corpus matching result,, named entity that mistake occurs further to the rule correction, and is input to rule set, the update rule collection with revised rule with correct named entity output;
F. constantly carry out rule correction according to labeled standards, up to the output result optimal, and definite optimum rule set.
2. the method for extracting Chinese named entity from text data according to claim 1 is characterized in that, the modification rule described in the d step comprises and merges rule, finger name extracting rule, border modification rule and type modification rule together.
3. the method for from text data, extracting Chinese named entity according to claim 2, it is characterized in that: described merging rule is about to should belong to an entity together among the Chinese word segmentation result and is that an entity merged in two or more words by false segmentation; The described name extracting rule that refers to together promptly finds the speech that refers to same name in the text and unifies mark; Described border modification rule is promptly revised the mistake of having lost self part when named entity extracts, if place name is lost suffix, then sets up corresponding place name suffix dictionary update information is provided; Described type modification rule, the type misjudgment when promptly revising the named entity extraction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910227302A CN101719122A (en) | 2009-12-04 | 2009-12-04 | Method for extracting Chinese named entity from text data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910227302A CN101719122A (en) | 2009-12-04 | 2009-12-04 | Method for extracting Chinese named entity from text data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101719122A true CN101719122A (en) | 2010-06-02 |
Family
ID=42433696
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910227302A Pending CN101719122A (en) | 2009-12-04 | 2009-12-04 | Method for extracting Chinese named entity from text data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101719122A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678262A (en) * | 2013-12-27 | 2014-03-26 | 中西亚通医疗信息科技（北京）有限公司 | Text processing method and text processing device |
| CN103678703A (en) * | 2013-12-30 | 2014-03-26 | 中国科学院自动化研究所 | Method and device for extracting open category named entity by means of random walking on map |
| CN104866472A (en) * | 2015-06-15 | 2015-08-26 | 百度在线网络技术（北京）有限公司 | Generation method and device of word segmentation training set |
| CN105068999A (en) * | 2015-08-14 | 2015-11-18 | 浪潮集团有限公司 | Method and apparatus for identifying amended entity words |
| CN105224622A (en) * | 2015-09-22 | 2016-01-06 | 中国搜索信息科技股份有限公司 | The place name address extraction of Internet and standardized method |
| CN105260360A (en) * | 2015-10-27 | 2016-01-20 | 小米科技有限责任公司 | Named entity identification method and device |
| CN106156017A (en) * | 2015-03-23 | 2016-11-23 | 北大方正集团有限公司 | Information identifying method and information identification system |
| CN106682220A (en) * | 2017-01-04 | 2017-05-17 | 华南理工大学 | Online traditional Chinese medicine text named entity identifying method based on deep learning |
| WO2017097166A1 (en) * | 2015-12-11 | 2017-06-15 | 北京国双科技有限公司 | Domain named entity recognition method and apparatus |
| CN107783961A (en) * | 2017-11-08 | 2018-03-09 | 郑州云海信息技术有限公司 | A kind of method, apparatus and readable storage medium storing program for executing of much-talked-about topic identification |
| CN109408825A (en) * | 2018-11-06 | 2019-03-01 | 杭州费尔斯通科技有限公司 | A kind of acceptance of the bid data extraction method based on name Entity recognition |
| CN109858018A (en) * | 2018-12-25 | 2019-06-07 | 中国科学院信息工程研究所 | A kind of entity recognition method and system towards threat information |
| CN110516252A (en) * | 2019-08-30 | 2019-11-29 | 京东方科技集团股份有限公司 | Data mask method, device, computer equipment and storage medium |
| CN111401059A (en) * | 2020-03-16 | 2020-07-10 | 深圳市子瑜杰恩科技有限公司 | Novel reading method |
| CN112115240A (en) * | 2019-06-21 | 2020-12-22 | 百度在线网络技术（北京）有限公司 | Classification processing method, classification processing device, server and storage medium |
| CN112445862A (en) * | 2020-11-27 | 2021-03-05 | 中国科学院信息工程研究所 | Internet of things equipment data set construction method and device, electronic equipment and storage medium |
| CN113190755A (en) * | 2021-05-11 | 2021-07-30 | 两比特(北京)科技有限公司 | Cloud system persona analysis system |
| CN113742498A (en) * | 2021-09-24 | 2021-12-03 | 国务院国有资产监督管理委员会研究中心 | Method for constructing and updating knowledge graph |
-
2009
- 2009-12-04 CN CN200910227302A patent/CN101719122A/en active Pending
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678262A (en) * | 2013-12-27 | 2014-03-26 | 中西亚通医疗信息科技（北京）有限公司 | Text processing method and text processing device |
| CN103678703B (en) * | 2013-12-30 | 2017-01-11 | 中国科学院自动化研究所 | Method and device for extracting open category named entity by means of random walking on map |
| CN103678703A (en) * | 2013-12-30 | 2014-03-26 | 中国科学院自动化研究所 | Method and device for extracting open category named entity by means of random walking on map |
| CN106156017A (en) * | 2015-03-23 | 2016-11-23 | 北大方正集团有限公司 | Information identifying method and information identification system |
| CN104866472A (en) * | 2015-06-15 | 2015-08-26 | 百度在线网络技术（北京）有限公司 | Generation method and device of word segmentation training set |
| CN105068999A (en) * | 2015-08-14 | 2015-11-18 | 浪潮集团有限公司 | Method and apparatus for identifying amended entity words |
| CN105224622A (en) * | 2015-09-22 | 2016-01-06 | 中国搜索信息科技股份有限公司 | The place name address extraction of Internet and standardized method |
| CN105260360B (en) * | 2015-10-27 | 2018-12-18 | 小米科技有限责任公司 | Name recognition methods and the device of entity |
| CN105260360A (en) * | 2015-10-27 | 2016-01-20 | 小米科技有限责任公司 | Named entity identification method and device |
| WO2017097166A1 (en) * | 2015-12-11 | 2017-06-15 | 北京国双科技有限公司 | Domain named entity recognition method and apparatus |
| CN106874256A (en) * | 2015-12-11 | 2017-06-20 | 北京国双科技有限公司 | Name the method and device of entity in identification field |
| CN106682220A (en) * | 2017-01-04 | 2017-05-17 | 华南理工大学 | Online traditional Chinese medicine text named entity identifying method based on deep learning |
| CN107783961A (en) * | 2017-11-08 | 2018-03-09 | 郑州云海信息技术有限公司 | A kind of method, apparatus and readable storage medium storing program for executing of much-talked-about topic identification |
| CN109408825A (en) * | 2018-11-06 | 2019-03-01 | 杭州费尔斯通科技有限公司 | A kind of acceptance of the bid data extraction method based on name Entity recognition |
| CN109858018A (en) * | 2018-12-25 | 2019-06-07 | 中国科学院信息工程研究所 | A kind of entity recognition method and system towards threat information |
| CN112115240A (en) * | 2019-06-21 | 2020-12-22 | 百度在线网络技术（北京）有限公司 | Classification processing method, classification processing device, server and storage medium |
| US11954439B2 (en) | 2019-08-30 | 2024-04-09 | Boe Technology Group Co., Ltd. | Data labeling method and device, and storage medium |
| CN110516252A (en) * | 2019-08-30 | 2019-11-29 | 京东方科技集团股份有限公司 | Data mask method, device, computer equipment and storage medium |
| WO2021036968A1 (en) * | 2019-08-30 | 2021-03-04 | 京东方科技集团股份有限公司 | Data labeling method and device, and storage medium |
| CN111401059A (en) * | 2020-03-16 | 2020-07-10 | 深圳市子瑜杰恩科技有限公司 | Novel reading method |
| CN112445862B (en) * | 2020-11-27 | 2024-01-26 | 中国科学院信息工程研究所 | Internet of things equipment data set construction method and device, electronic equipment and storage medium |
| CN112445862A (en) * | 2020-11-27 | 2021-03-05 | 中国科学院信息工程研究所 | Internet of things equipment data set construction method and device, electronic equipment and storage medium |
| CN113190755A (en) * | 2021-05-11 | 2021-07-30 | 两比特(北京)科技有限公司 | Cloud system persona analysis system |
| CN113742498A (en) * | 2021-09-24 | 2021-12-03 | 国务院国有资产监督管理委员会研究中心 | Method for constructing and updating knowledge graph |
| CN113742498B (en) * | 2021-09-24 | 2024-04-09 | 国务院国有资产监督管理委员会研究中心 | Knowledge graph construction and updating method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101719122A (en) | Method for extracting Chinese named entity from text data | |
| CN103123618B (en) | Text similarity acquisition methods and device | |
| US20150186361A1 (en) | Method and apparatus for improving a bilingual corpus, machine translation method and apparatus | |
| CN105095190B (en) | A kind of sentiment analysis method combined based on Chinese semantic structure and subdivision dictionary | |
| CN103077164A (en) | Text analysis method and text analyzer | |
| CN110929520B (en) | Unnamed entity object extraction method and device, electronic equipment and storage medium | |
| Darwish et al. | Using Stem-Templates to Improve Arabic POS and Gender/Number Tagging. | |
| CN106843840A (en) | A kind of version evolving annotation multiplexing method of source code based on similarity analysis | |
| CN110188359B (en) | Text entity extraction method | |
| CN108647511A (en) | The password strength assessment method derived based on weak passwurd | |
| CN109977398A (en) | A kind of speech recognition text error correction method of specific area | |
| CN109766556B (en) | Corpus restoration method and device | |
| Jayan et al. | A hybrid statistical approach for named entity recognition for malayalam language | |
| CN105183742A (en) | Resume identification method | |
| WO2023116561A1 (en) | Entity extraction method and apparatus, and electronic device and storage medium | |
| CN106383814A (en) | Word segmentation method of English social media short text | |
| CN106681716A (en) | Intelligent terminal and automatic classification method of application programs thereof | |
| CN112926313A (en) | Method and system for extracting slot position information | |
| CN104598441B (en) | A kind of method that computer splits Chinese sentence | |
| CN108268669A (en) | A kind of crucial new word discovery method based on multidimensional words and phrases feature and sentiment analysis | |
| CN104239292B (en) | A kind of method for obtaining specialized vocabulary translation | |
| Huang et al. | A realistic and robust model for Chinese word segmentation | |
| CN106126497A (en) | A kind of automatic mining correspondence executes leader section and the method for cited literature textual content fragment | |
| CN111339314B (en) | Ternary group data generation method and device and electronic equipment | |
| CN107315739A (en) | A kind of semantic analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100602 |