CN102033950A - Construction method and identification method of automatic electronic product named entity identification system - Google Patents
Construction method and identification method of automatic electronic product named entity identification system Download PDFInfo
- Publication number
- CN102033950A CN102033950A CN 201010602773 CN201010602773A CN102033950A CN 102033950 A CN102033950 A CN 102033950A CN 201010602773 CN201010602773 CN 201010602773 CN 201010602773 A CN201010602773 A CN 201010602773A CN 102033950 A CN102033950 A CN 102033950A
- Authority
- CN
- China
- Prior art keywords
- electronic product
- named entity
- product named
- identification
- feature
- 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 construction method and an identification method of an automatic electronic product named entity identification system, relates to a construction method and an identification method of a named entity identification system in natural language processing and belongs to a technique for automatically identifying names of electronic products from related information. The invention is used for identifying the names of the electronic products and solves the problems that a rule-based identification system has low recall rate during identification and a machine learning-based identification system needs to manually label a great deal of training language database during identification. The construction method comprises the following steps of: forming a knowledge base of the original linguistic data; constructing a label language database; and performing electronic product named entity identification on the basis of a conditional random field method. The identification method comprises the following steps that: a free text is input into the automatic electronic product named entity identification system; and the system extracts characteristics by using a characteristic template, acquires each weight corresponding to each characteristic by using a conditional random field model and calculates the weights by the conditional random field method to acquire an identification result.
Description
Technical field
The present invention relates to the construction method and the recognition methods of the named entity recognition system in the natural language processing, belong to the technology of from relevant information, the title of electronic product being discerned automatically.
Background technology
Outwardness and the things that can distinguish mutually are referred to as entity.Entity can be concrete people, thing and thing, also can be abstract concept or contact.The named entity recognition task is meant the entity that has certain sense in the identification text.Along with human society steps into digital Age, increasing electronic product has entered into people's life.Various reports about electronic product appear in the electronic document in a large number.Advertisement, using method and user comment have been full of on the internet especially about electronic product.Electronic product named entity recognition technology can help people better inquiry and the own interested electronic product information of management, help enterprise to find interconnected user on the network quickly to the feedback of own product and deliver advertisement more accurately, so this technology more and more receive the concern of industry member and academia.
Present named entity recognition technology is primarily aimed at these traditional named entities such as name in the news language material, place name, mechanism's name.Major technology can be divided into two classes: rule-based technology and based on the technology of machine learning.Rule-based technology is mainly utilized the composition rule of named entity, adopts the mode of artificial constructed knowledge base and rule base to carry out Entity recognition, and this technology accuracy rate is higher, but recall rate is lower, and is difficult to transplant.Technology based on machine learning mainly adopts machine learning algorithm and contextual feature to carry out Entity recognition, and wherein important recognizer comprises Hidden Markov Model (HMM), maximum entropy model, supporting vector machine model, decision-tree model etc.These class methods need a large amount of corpus of artificial mark, for lacking the common poor effect of electronic product named entity that marks language material.
Compare with traditional named entity, the electronic product named entity has following characteristics: it is faster 1) to upgrade variation; 2) constitute complicatedly, and mix usually and a large amount of numerals is arranged and stride language character; 3) lack standard and mark language material.The research for the electronic product named entity recognition at present both at home and abroad still is in the starting stage, and each sticks to his own version or argument to the definition of electronic product named entity.And corresponding recognition method also mainly concentrates on and directly applies mechanically traditional named entity recognition technology, lacks specific aim, so recognition accuracy and recall rate all are difficult to reach realistic scale.
Summary of the invention
The construction method that the purpose of this invention is to provide a kind of electronic product named entity automatic recognition system, with solve rule-based recognition system when identification recall rate lower, and need the problem of a large amount of corpus of artificial mark during based on the recognition system identification of machine learning.
It comprises the steps: one, utilizes downloaded software to collect the electronic product info web of multiple type from the internet, extracts the text of info web, thereby forms the knowledge base of original language material; Use participle part-of-speech tagging instrument, original language material is carried out participle and part-of-speech tagging processing, according to the definition of electronic product named entity, the language material behind participle and the part-of-speech tagging is carried out the entity mark afterwards, make up a tagged corpus; Described definition to the electronic product named entity is meant that brand name according to an electronic product named entity, series name and model three parts distinguish the electronic product named entity; Two, based on condition random territory method, define a plurality of feature templates, feature templates utilization mark language material and knowledge base specifically dissolve feature, the operation result of condition random territory method on tagged corpus can be given certain weight for each feature, and the condition random domain model that feature and its corresponding weight constitute just can be used for carrying out the electronic product named entity recognition.
The present invention also provides the recognition methods based on above-mentioned electronic product named entity automatic recognition system, and it comprises the steps: one, the free text that is used to discern is imported described electronic product named entity automatic recognition system; Two, system at first utilizes feature templates to extract feature, utilizes the condition random domain model to obtain the weight of each feature correspondence then, utilizes these weights condition random territory method to carry out computing and just obtains final recognition result.
Method of the present invention uses participle part-of-speech tagging instrument to come the electronic product info web of collecting in the internet is handled, avoided by a large amount of corpus of artificial mark, handle free text based on condition random territory method and knowledge base, tagged corpus, so recognition system recall rate height when identification.The commercial matters information of magnanimity on the internet can be managed and organize to method of the present invention effectively, and the raising people search, the efficient of management and use information.The present invention utilizes the composing law of electronic product named entity, change characteristics fast, various informativeization at product class named entity, electronic product named entity recognition method based on the condition random domain model of knowledge base has been proposed, and having realized corresponding system, the accuracy rate and the recall rate of system identification all reach more than 86%.
Description of drawings
Fig. 1 is the schematic flow sheet of embodiment of the present invention five, and Fig. 2 is the example schematic of an identification in the embodiment five.
Embodiment
Embodiment one: the construction method of the electronic product named entity automatic recognition system of present embodiment comprises the steps: one, utilizes downloaded software to collect the electronic product info web of multiple type from the internet, extract the text of info web, thereby form the knowledge base of original language material; Use participle part-of-speech tagging instrument, original language material is carried out participle (spliting between speech in the sentence and speech) and part-of-speech tagging processing (marking the part of speech of each speech), afterwards according to the definition of electronic product named entity, language material behind participle and the part-of-speech tagging is carried out the entity mark, make up a tagged corpus; Described definition to the electronic product named entity is meant that brand name according to an electronic product named entity, series name and model three parts distinguish the electronic product named entity; Two, based on condition random territory method, define a plurality of feature templates, feature templates utilization mark language material and knowledge base specifically dissolve feature, the operation result of condition random territory method on tagged corpus can be given certain weight for each feature, and the condition random domain model that feature and its corresponding weight constitute just can be used for carrying out the electronic product named entity recognition.
Embodiment two: the difference of present embodiment and embodiment one is: the resource in the knowledge base all utilizes web crawlers technology and information extraction technique to obtain automatically from the internet; Described knowledge base comprises: have the brand name dictionary that the brand message characteristic is constructed at electronic product; The series name dictionary that the branch structure of series is arranged at the electronic product under the brand; Or has the particular words knowledge base that the phrase of certain sense is constructed at some.
Embodiment three: the recognition methods based on the electronic product named entity automatic recognition system of embodiment one of present embodiment, it comprises the steps: one, the free text that is used to discern is imported described electronic product named entity automatic recognition system; Two, system at first utilizes feature templates to extract feature, utilizes the condition random domain model to obtain the weight of each feature correspondence then, utilizes these weights condition random territory method to carry out computing and just obtains final recognition result.
Embodiment four: the difference of present embodiment and embodiment three is: it also comprises step 3, adopts regular modification method that the electronic product named entity after discerning is revised, and described modification rule is by obtaining based on wrong method of driving.
Embodiment five: below in conjunction with Fig. 1 and Fig. 2, technical solution of the present invention is clearly and completely described, obviously, described embodiment only is the present invention's part embodiment, rather than whole embodiment.Based on the embodiment among the present invention, those of ordinary skills belong to the scope of protection of the invention not making all other embodiment that obtained under the creative work prerequisite.
Present embodiment provides the automatically method of identification of a kind of electronic product named entity, can find automatically with the free text of recognition network in the associated electrical product naming entity.Specify embodiments of the present invention below in conjunction with Fig. 1.Present embodiment comprises: the 1) structure of corpus and knowledge base; 2) structure of feature extraction and feature templates; 3) will be applied in the electronic product named entity recognition based on the method that the machine learning and the rule of knowledge base are revised.
(1) definition of electronic product named entity and corpus make up.
In the present invention,, product naming entity is divided into three parts according to the difference of product component renewal frequency, i.e. the brand of product (BRA), series (SER), model (TYP), and product naming entity is described attribute by these three assemblies and some and is formed.Wherein, brand refers to intrinsic noun---the trade mark of product, for example " Nokia "; Series name is meant a series of under the brand, and for example " EasyShare " is exactly a series under Kodak's digital camera brand; Model name refers to the version information under brand or the series, is made up of letter, numeral and some symbols, and for example " Nokia N70 " middle N70 is exactly a model under the Nokia product brand.Usually, the highest part of renewal frequency is a product type in the digital product field, and in definition of the present invention, model name will be discerned as an independent classification, after the method that combines statistics and rule, this classification can effectively be discerned, and has higher renewal tolerance and accuracy of identification.
In addition, in actual applications, brand name and exabyte are born the same name, and this may produce ambiguity.For example " Nokia has released the mobile phone of a novel fashion recently---the N70 of Nokia.”。First " Nokia " refers to Nokia in the statement, and second " Nokia " then represents brand.The brand name that also has not only can be born the same name with exabyte, and under particular environment, it represents other classifications, and for example " apple " can be brand name, exabyte, also can be fruit, and this just need determine the implication that such speech is correct by contextual environment.That is to say to have only the name part of speech structure of working as to contain following determinacy product information, just can constitute product naming entity.
1) contain product brand, series or model entity any one, two or three, as " Nokia 5800 ", " 6300 mobile phone " is a ProductName entity, " Nokia " is the brand name entity, " 6300 " are the model name entities, and " digital camera " then is not a ProductName entity;
2) although do not contain brand, series or type information, contain peculiar product another name of certain brand or version information, be a ProductName entity as " little black notebook ", because " little black " is the another name of Thinkpad series under the association.And " intelligent " is the attribute information that all brands can have in " intelligent mobile phone ", so it is not a ProductName entity.
On the basis of above-mentioned electronic product named entity definition, utilize web crawlers from the internet, to collect the electronic product network information of multiple type, the text that extracts info web is built into original corpus.Utilize brand, serial knowledge base and participle part-of-speech tagging instrument then, original language material is carried out participle and part-of-speech tagging processing, according to the definition of electronic product named entity, the language material behind participle and the part-of-speech tagging is carried out the entity mark afterwards.After processing is handled to original corpus, use correlation technique that the processing corpus is carried out the consistance evaluation and test again, language material not up to standard is marked again, acquire a certain degree up to evaluation metrics.Through above-mentioned steps, made up one contain much information, standard corpus that the text type is various.
(2) based on the electronic product named entity recognition method of the condition random domain model of knowledge base.
Named entity recognition is extremely important for the aftertreatment of text message.In named entity recognition, the most frequently used machine learning method has maximum entropy model, the latent markov model of maximum entropy and condition random domain model.This three class model has closely similar common ground, belongs to the discriminant model.Used the best condition random domain model of effect among the present invention.Machine learning model all needs to make up a feature templates, is used for extracting when the current speech of identification the contextual feature of this speech.Therefore, the structure of feature templates is particularly important.
In the present invention, the process of ProductName Entity recognition is seen as a polytypic process, the target classification that wherein needs to discern comprises product brand name, series name, model name, reaches exabyte and product entity, the name entity of each classification is subdivided into beginning part, center section and whole three kinds of situations again, all is classified as a class not belonging to word any in the above classification.In order to solve the difficult problem that triggers of candidate's entity in the product naming entity identification, introduced product brand storehouse and serial storehouse trigger condition, simultaneously the identification of the triggering model name entity that they also can be covert as brand entity and serial entity.Feature in the model is produced by a series of feature templates.Defined 13 monobasic feature templates altogether, utilized the mode of these 13 monobasic feature templates that define then, be built into the set of binary feature template, carried out the screening of binary feature template then with information gain by making up in twos.Screening is made up in twos with monobasic feature templates and binary feature template after finishing, and is built into the ternary feature templates, uses the same method and screens, and has screened more than 40 feature templates construction feature template set at last.
After setting up the feature templates set, from the standard corpus of mark, select a certain proportion of language material composing training set at random, utilization feature templates collection and machine learning algorithm are trained, and by the adjustment to frequency of training, make the good match actual text of model energy that trains.Use the rule learning algorithm again, analyze the experimental result mistake, extract the rule that corrects mistakes, the improvement system is to the performance of electronic product named entity recognition.
More than electronic product named entity automatic identifying method and system that the embodiment of the invention provided are described in detail, used specific case herein principle of the present invention and embodiment are set forth, the explanation of above embodiment just is used for helping to understand method of the present invention and core concept thereof; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, the part that all can change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.
Claims (4)
1. the construction method of electronic product named entity automatic recognition system, it is characterized in that it comprises the steps: one, utilizes downloaded software to collect the electronic product info web of multiple type from the internet, extract the text of info web, thereby form the knowledge base of original language material; Use participle part-of-speech tagging instrument, original language material is carried out participle and part-of-speech tagging processing, according to the definition of electronic product named entity, the language material behind participle and the part-of-speech tagging is carried out the entity mark afterwards, make up a tagged corpus; Described definition to the electronic product named entity is meant that brand name according to an electronic product named entity, series name and model three parts distinguish the electronic product named entity; Two, based on condition random territory method, define a plurality of feature templates, feature templates utilization mark language material and knowledge base specifically dissolve feature, the operation result of condition random territory method on tagged corpus can be given certain weight for each feature, and the condition random domain model that feature and its corresponding weight constitute just can be used for carrying out the electronic product named entity recognition.
2. the construction method of electronic product named entity automatic recognition system according to claim 1 is characterized in that the resource in the knowledge base all utilizes web crawlers technology and information extraction technique to obtain automatically from the internet; Described knowledge base comprises: have the brand name dictionary that the brand message characteristic is constructed at electronic product; The series name dictionary that the branch structure of series is arranged at the electronic product under the brand; Or has the particular words knowledge base that the phrase of certain sense is constructed at some.
3. the recognition methods of the electronic product named entity automatic recognition system that makes up based on the described method of claim 1 is characterized in that it comprises the steps: one, the free text that is used to discern is imported described electronic product named entity automatic recognition system; Two, system at first utilizes feature templates to extract feature, utilizes the condition random domain model to obtain the weight of each feature correspondence then, utilizes these weights condition random territory method to carry out computing and just obtains final recognition result.
4. electronic product named entity automatic identifying method according to claim 3, it is characterized in that it also comprises step 3, adopts regular modification method that the electronic product named entity after discerning is revised, described modification rule is by obtaining based on wrong method of driving.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010602773 CN102033950A (en) | 2010-12-23 | 2010-12-23 | Construction method and identification method of automatic electronic product named entity identification system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201010602773 CN102033950A (en) | 2010-12-23 | 2010-12-23 | Construction method and identification method of automatic electronic product named entity identification system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102033950A true CN102033950A (en) | 2011-04-27 |
Family
ID=43886843
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201010602773 Pending CN102033950A (en) | 2010-12-23 | 2010-12-23 | Construction method and identification method of automatic electronic product named entity identification system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102033950A (en) |
Cited By (24)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609853A (en) * | 2012-02-27 | 2012-07-25 | 蒋永 | System and method for intelligently identifying names and models of commodities |
CN103034693A (en) * | 2012-12-03 | 2013-04-10 | 哈尔滨工业大学 | Open-type entity and type identification method thereof |
CN103164426A (en) * | 2011-12-13 | 2013-06-19 | 北大方正集团有限公司 | Method and device of recognizing named entity |
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 |
CN104331395A (en) * | 2014-10-28 | 2015-02-04 | 北京京东尚科信息技术有限公司 | Method and device for identifying Chinese product name from text |
CN104572625A (en) * | 2015-01-21 | 2015-04-29 | 北京云知声信息技术有限公司 | Recognition method of named entity |
CN105426358A (en) * | 2015-11-09 | 2016-03-23 | 中国农业大学 | Automatic disease noun identification method |
CN105574111A (en) * | 2015-12-10 | 2016-05-11 | 天津海量信息技术有限公司 | Enterprise entity authentication method based on enterprise attribute library |
CN105812473A (en) * | 2016-03-29 | 2016-07-27 | 成都小多科技有限公司 | Data processing method and device |
CN105868193A (en) * | 2015-01-19 | 2016-08-17 | 富士通株式会社 | Device and method used to detect product relevant information in electronic text |
TWI554896B (en) * | 2012-07-30 | 2016-10-21 | Alibaba Group Services Ltd | Information Classification Method and Information Classification System Based on Product Identification |
CN106407183A (en) * | 2016-09-28 | 2017-02-15 | 医渡云(北京)技术有限公司 | Method and device for generating medical named entity recognition system |
WO2017028422A1 (en) * | 2015-08-20 | 2017-02-23 | 小米科技有限责任公司 | Knowledge base construction method and apparatus |
CN103678336B (en) * | 2012-09-05 | 2017-04-12 | 阿里巴巴集团控股有限公司 | Method and device for identifying entity words |
CN106776560A (en) * | 2016-12-15 | 2017-05-31 | 昆明理工大学 | A kind of Kampuchean organization name recognition method |
CN106897559A (en) * | 2017-02-24 | 2017-06-27 | 黑龙江特士信息技术有限公司 | A kind of symptom and sign class entity recognition method and device towards multi-data source |
CN106909656A (en) * | 2017-02-27 | 2017-06-30 | 腾讯科技(深圳)有限公司 | Obtain the method and device of Text Feature Extraction model |
CN107423264A (en) * | 2017-07-10 | 2017-12-01 | 广东华联建设投资管理股份有限公司 | A kind of engineering material borrowing-word extracting method |
CN108009229A (en) * | 2017-11-29 | 2018-05-08 | 厦门市美亚柏科信息股份有限公司 | Method, terminal device and the storage medium that public sentiment event data is found |
CN108108344A (en) * | 2016-11-24 | 2018-06-01 | 北京智能管家科技有限公司 | A kind of method and device for combining identification and connection name entity |
CN110046978A (en) * | 2019-03-19 | 2019-07-23 | 上海大学 | Intelligent method of charging out |
CN110704633A (en) * | 2019-09-04 | 2020-01-17 | 平安科技(深圳)有限公司 | Named entity recognition method and device, computer equipment and storage medium |
CN112818693A (en) * | 2021-02-07 | 2021-05-18 | 深圳市世强元件网络有限公司 | Automatic extraction method and system for electronic component model words |
CN114297990A (en) * | 2021-12-22 | 2022-04-08 | 北京捷通华声科技股份有限公司 | Public opinion monitoring method and device, computer readable storage medium and processor |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101261635A (en) * | 2008-04-29 | 2008-09-10 | 哈尔滨工业大学深圳研究生院 | Passive type network information automatic highly effective collection system and method |
-
2010
- 2010-12-23 CN CN 201010602773 patent/CN102033950A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101261635A (en) * | 2008-04-29 | 2008-09-10 | 哈尔滨工业大学深圳研究生院 | Passive type network information automatic highly effective collection system and method |
Non-Patent Citations (1)
Title |
---|
《郑州大学学报(理学版)》 20100331 梅丰等 面向网络文本的中文产品命名实体识别 论文62页倒数第7行-65页倒数第11行,表1-3 1-4 第42卷, 第1期 2 * |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103164426A (en) * | 2011-12-13 | 2013-06-19 | 北大方正集团有限公司 | Method and device of recognizing named entity |
CN103164426B (en) * | 2011-12-13 | 2015-10-28 | 北大方正集团有限公司 | A kind of method of named entity recognition and device |
CN102609853A (en) * | 2012-02-27 | 2012-07-25 | 蒋永 | System and method for intelligently identifying names and models of commodities |
TWI554896B (en) * | 2012-07-30 | 2016-10-21 | Alibaba Group Services Ltd | Information Classification Method and Information Classification System Based on Product Identification |
CN103678336B (en) * | 2012-09-05 | 2017-04-12 | 阿里巴巴集团控股有限公司 | Method and device for identifying entity words |
CN103034693B (en) * | 2012-12-03 | 2016-03-02 | 哈尔滨工业大学 | Open entity and kind identification method thereof |
CN103034693A (en) * | 2012-12-03 | 2013-04-10 | 哈尔滨工业大学 | Open-type entity and type identification method thereof |
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 |
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 |
CN104331395B (en) * | 2014-10-28 | 2017-11-03 | 北京京东尚科信息技术有限公司 | The method and apparatus that Chinese trade name is recognized from text |
CN104331395A (en) * | 2014-10-28 | 2015-02-04 | 北京京东尚科信息技术有限公司 | Method and device for identifying Chinese product name from text |
CN105868193A (en) * | 2015-01-19 | 2016-08-17 | 富士通株式会社 | Device and method used to detect product relevant information in electronic text |
CN104572625A (en) * | 2015-01-21 | 2015-04-29 | 北京云知声信息技术有限公司 | Recognition method of named entity |
US10331648B2 (en) | 2015-08-20 | 2019-06-25 | Xiaomi Inc. | Method, device and medium for knowledge base construction |
WO2017028422A1 (en) * | 2015-08-20 | 2017-02-23 | 小米科技有限责任公司 | Knowledge base construction method and apparatus |
CN105426358A (en) * | 2015-11-09 | 2016-03-23 | 中国农业大学 | Automatic disease noun identification method |
CN105426358B (en) * | 2015-11-09 | 2018-08-31 | 中国农业大学 | A kind of disease noun automatic identifying method for magnanimity news |
CN105574111A (en) * | 2015-12-10 | 2016-05-11 | 天津海量信息技术有限公司 | Enterprise entity authentication method based on enterprise attribute library |
CN105812473A (en) * | 2016-03-29 | 2016-07-27 | 成都小多科技有限公司 | Data processing method and device |
CN105812473B (en) * | 2016-03-29 | 2020-01-17 | 成都晓多科技有限公司 | Data processing method and device |
CN106407183A (en) * | 2016-09-28 | 2017-02-15 | 医渡云(北京)技术有限公司 | Method and device for generating medical named entity recognition system |
CN106407183B (en) * | 2016-09-28 | 2019-06-28 | 医渡云(北京)技术有限公司 | Medical treatment name entity recognition system generation method and device |
CN108108344A (en) * | 2016-11-24 | 2018-06-01 | 北京智能管家科技有限公司 | A kind of method and device for combining identification and connection name entity |
CN106776560A (en) * | 2016-12-15 | 2017-05-31 | 昆明理工大学 | A kind of Kampuchean organization name recognition method |
CN106897559A (en) * | 2017-02-24 | 2017-06-27 | 黑龙江特士信息技术有限公司 | A kind of symptom and sign class entity recognition method and device towards multi-data source |
CN106909656A (en) * | 2017-02-27 | 2017-06-30 | 腾讯科技(深圳)有限公司 | Obtain the method and device of Text Feature Extraction model |
CN107423264A (en) * | 2017-07-10 | 2017-12-01 | 广东华联建设投资管理股份有限公司 | A kind of engineering material borrowing-word extracting method |
CN108009229A (en) * | 2017-11-29 | 2018-05-08 | 厦门市美亚柏科信息股份有限公司 | Method, terminal device and the storage medium that public sentiment event data is found |
CN110046978A (en) * | 2019-03-19 | 2019-07-23 | 上海大学 | Intelligent method of charging out |
CN110704633A (en) * | 2019-09-04 | 2020-01-17 | 平安科技(深圳)有限公司 | Named entity recognition method and device, computer equipment and storage medium |
CN110704633B (en) * | 2019-09-04 | 2023-07-21 | 平安科技(深圳)有限公司 | Named entity recognition method, named entity recognition device, named entity recognition computer equipment and named entity recognition storage medium |
CN112818693A (en) * | 2021-02-07 | 2021-05-18 | 深圳市世强元件网络有限公司 | Automatic extraction method and system for electronic component model words |
CN114297990A (en) * | 2021-12-22 | 2022-04-08 | 北京捷通华声科技股份有限公司 | Public opinion monitoring method and device, computer readable storage medium and processor |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102033950A (en) | Construction method and identification method of automatic electronic product named entity identification system | |
CN106649818B (en) | Application search intention identification method and device, application search method and server | |
CN104268197B (en) | A kind of industry comment data fine granularity sentiment analysis method | |
CN104572958B (en) | A kind of sensitive information monitoring method based on event extraction | |
CN107729468B (en) | answer extraction method and system based on deep learning | |
CN108664599B (en) | Intelligent question-answering method and device, intelligent question-answering server and storage medium | |
CN107169079B (en) | A kind of field text knowledge abstracting method based on Deepdive | |
CN105260356A (en) | Chinese interactive text emotion and topic identification method based on multitask learning | |
CN112184525A (en) | System and method for realizing intelligent matching recommendation through natural semantic analysis | |
CN113505200B (en) | Sentence-level Chinese event detection method combined with document key information | |
CN103049435A (en) | Text fine granularity sentiment analysis method and text fine granularity sentiment analysis device | |
Layton et al. | Recentred local profiles for authorship attribution | |
CN101127042A (en) | Sensibility classification method based on language model | |
CN105183715B (en) | A kind of word-based distribution and the comment spam automatic classification method of file characteristics | |
CN108038208B (en) | Training method and device of context information recognition model and storage medium | |
CN103020293A (en) | Method and system for constructing ontology base in mobile application | |
CN104881458A (en) | Labeling method and device for web page topics | |
CN102662923A (en) | Entity instance leading method based on machine learning | |
CN109145301B (en) | Information classification method and device and computer readable storage medium | |
CN110910175B (en) | Image generation method for travel ticket product | |
CN108388914A (en) | A kind of grader construction method, grader based on semantic computation | |
CN104346326A (en) | Method and device for determining emotional characteristics of emotional texts | |
CN102708164A (en) | Method and system for calculating movie expectation | |
CN106202039B (en) | Vietnamese portmanteau word disambiguation method based on condition random field | |
CN103886020A (en) | Quick search method of real estate information |
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 |
Application publication date: 20110427 |