CN109670045A - Emotion reason abstracting method based on ontology model and multi-kernel support vector machine - Google Patents
Emotion reason abstracting method based on ontology model and multi-kernel support vector machine Download PDFInfo
- Publication number
- CN109670045A CN109670045A CN201811303734.1A CN201811303734A CN109670045A CN 109670045 A CN109670045 A CN 109670045A CN 201811303734 A CN201811303734 A CN 201811303734A CN 109670045 A CN109670045 A CN 109670045A
- Authority
- CN
- China
- Prior art keywords
- emotion
- emotion reason
- reason
- ontology model
- event
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/242—Dictionaries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Machine Translation (AREA)
Abstract
The present invention discloses a kind of emotion reason abstracting method based on ontology model and multi-kernel support vector machine, this method key step are as follows: building emotion reason corpus first；Then emotion reason ontology model is constructed according to ABC ontology model；Emotion reason event is identified based on algorithm of support vector machine on this model.The invention proposes emotion reason ontology model is based on, which expands base matrix model by the relationship in fusion emotion reason field between emotional event, emotion reason, people, region, behavioral concept and concept based on ABC ontology.Then it constructs based on emotion reason tagged corpus, defines the expression of emotion reason event formization, realize the emotion reason event recognition algorithm based on algorithm of support vector machine.The extraction of emotion reason mainly is carried out in emotion reason ontology model, it is found that the causality in text information, the method have a wide range of applications in the prediction of the prediction of event, affair clustering and stock market.
Description
Technical field
The invention belongs to natural language processing fields, and in particular to a kind of based on ontology model and multi-kernel support vector machine
Emotion reason abstracting method proposes emotion reason ontology model, is then carried out on this model using multi-kernel support vector machine
Emotion reason extracts, to realize the trigger event for excavating that excitation emotion is generated and shifted in text information.
Background technique
With the rapid growth of social network-i i-platform, more and more people tend to express their feelings on social networks
Sense, the correlative study that text emotion calculates at present mainly include mood analysis, the identification of mood reason, emotional prediction etc., Chinese
Earliest, most study, the research that emotion reason extracts is started late for the research starting of this sentiment analysis.But extract emotion reason
Feelings monitoring is had far-reaching significance, for example, the reason of facing emergency event, wondering the emotion of the common people and generate emotion, according to
According to secondary and then understand public sentiment trend；Film making quotient wonders the emotion and its emotion reason of viewing people, to produce
The film that viewing people prefers.This part is mainly the present Research for introducing the excavation of emotion reason.
SophiaM.Y.Lee[1]It is proposed that emotion reason excavates this concept, the i.e. event of triggering mood generation, tool for the first time
Body shows as certain situation occurred, certain situation occurred or specific object, can be instantaneous, is also possible to duration
's.Relevant research approach is also to be driven by linguistic rules, extracts its correspondence for the emotion expression service in newsletter archive
The reason of.The problem of with going deep into for research, starting some methods based on machine learning model occur, emotion reason extracted
It is converted into the classification problem of reason candidate.Chen[2]The detection of mood reason is realized etc. a kind of multi-tag method is proposed, this
Kind method not only can be found that the reason of across clause, moreover it is possible to the reason of providing useful long range information.Gui[3]Deng attempting for the first time
Mood cause information in microblogging text is excavated, rule-based method and the method based on machine learning has been respectively adopted.
Bibliography:
[1]Sophia Yat Mei Lee,YingChen,and Chu-Ren Huang.2010.A text-driven
rule-based system for emotioncause detection.In Coling 2010,Beijing,China.
[2]Chen Y,Lee S Y M,Li S,et al.Emotion Cause Detection with
Linguistic Constructions[C]//Proceedings of the 23rd International Conference
on Computational Linguistics,Beijing,2010:179-187.
[3]Lin Gui,Dongyin Wu,YuZhou,Qin Lu and RuifengXu.2016.Event-Driven
Emotion Cause Extraction withCorpus Construction,in Proceedings of Empirical
Methods for Natural LanguageProcessing(EMNLP),pp.1639-1649.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, propose a kind of based on ontology model and multicore supporting vector
The emotion reason abstracting method of machine.
The present invention is to solve the technical issues of proposing in background technique, and used technical solution is as follows: being based on ontology mould
The emotion reason abstracting method of type and multi-kernel support vector machine, this method key step are as follows: building emotion reason corpus first,
Then emotion reason ontology model is constructed according to ABC ontology model, finally on this model based on algorithm of support vector machine to feelings
Sense reason event is identified.
The method specifically includes following steps:
1) crawler code to be write, crawls news data, each news data content includes model title, model content,
Model is delivered the time, news url, specific building such as Fig. 1；
2) data are pre-processed, removes stop words, segmented；
3) emotion reason corpus is constructed；
4) emotion reason ontology model, specific building such as Fig. 2 are constructed according to ABC ontology model；
5) by the concepts such as emotional event, emotion reason, people, region, behavior in fusion emotion reason field and concept it
Between relationship expand emotion reason ontology model；
6) emotion reason tree is constructed according to ternary relation；
7) judge whether read statement contains emotion word；
8) emotion reason event is extracted using algorithm of support vector machine；
9) accurate rate P value is used, recall rate R value and F value assess emotional event extraction algorithm；
The formula for calculating accuracy rate is as follows:
The formula for calculating recall rate is as follows:
The formula for calculating F value is as follows:
10) emotion reason is analyzed on emotion reason ontology model；
Step 1) of the present invention is mainly using JAVA language under the environmental level of myeclipse, using including
Struts, Spring, Hibernate framework technology are realized.
Data preprocessing operation, pre-treatment step packet are carried out with existing Chinese word segmentation tool in step 2) of the present invention
It includes and removes stop words, punctuate, participle, part-of-speech tagging or semantic analysis.
Punctuate processing of the present invention is mainly realized using existing segmenter.
Segmenter of the present invention mainly includes word segmenter, Ansj segmenter, Stamford participle, Lucene&Nutch
Segmenter, Stamford segmenter or Lucene&Nutch segmenter, Hanlp, Jieba.
Step 4) emotion reason ontology model of the present invention is mainly according to preceding because of-behavior-consequence ABC ontology model
Constructed, ABC ontology model is mainly modeled according to event, the content of event mainly pass through agent, action,
Situation, event, place description, such as Fig. 2.
Step 5) of the present invention is specifically: being extended region such as Fig. 3, to emotion and is extended such as Fig. 4.
The utility model has the advantages that
The method that the prior art excavates emotion cause and effect is not bound with related fields information first, this time uses letter
The method of single scale template, event matches effect is also barely satisfactory.Compared with prior art, method proposes be based on feelings
Feel reason ontology model, the model based on ABC ontology, by fusion emotion reason field in emotional event, emotion reason,
Relationship between people, region, behavioral concept and concept expands base matrix model.Then it constructs based on emotion reason
Tagged corpus defines the expression of emotion reason event formization, realizes the emotion reason thing based on algorithm of support vector machine
Part recognizer.The extraction of emotion reason mainly is carried out in emotion reason ontology model, finds the causality in text information,
The method has a wide range of applications in the prediction of the prediction of event, affair clustering and stock market.Therefore, this method mainly has
It is following
The utility model has the advantages that
1, method proposes emotion reason ontology models, merge pertinent arts, are more advantageous to and excavate implicit feelings
Feel reason.
2, this method constructs the emotion reason corpus of related fields.
3, the emotion reason abstracting method based on support vector machines is proposed using the method for machine learning, the method is than it
Preceding rule-based emotion reason abstracting method has preferable performance in accuracy rate, recall rate, F value.
4, this method has good scalability, cannot only be applied to emotion reason excavation applications, but also can apply
To any reason excavation applications.
Detailed description of the invention
Fig. 1 is news corpus library example.
Fig. 2 is emotion reason ontology model.
Fig. 3 is that region is extended model.
Fig. 4 is emotion extended model.
Fig. 5 is the emotion reason abstracting method flow chart based on ontology model and multi-kernel support vector machine.
Specific embodiment
Below in conjunction with the drawings and specific embodiments, the present invention will be further described in detail.
Embodiment 1
The proposed in this paper kind of emotion reason abstracting method based on ontology model and multi-kernel support vector machine, mainly finds
Emotion reason in text can be used for understanding the public sentiment trend of grave news event, and specific embodiment is as follows:
Fig. 1 is the emotion reason abstracting method flow chart based on ontology model and multi-kernel support vector machine.
Step 1: using JAVA language under the environmental level of myeclipse, using Struts, Spring,
The framework technologies such as Hibernate carry out realizing a crawler frame, crawl the news data of Sina's " 812 Tianjin explosive incident ",
Input data as this method.The method that crawls mainly is divided into three steps, first according to kind of a sublink, extracts Object linking and is put into
Queue to be crawled；Then the information that needs are parsed and extracted from the page, finally handles data, by the data extracted with
File format storage or deposit database and search engine index library etc..
Step 2: the text data to the ends of the earth pre-processes, in the process, present case mainly chooses grave news
Then event makes pauses in reading unpunctuated ancient writings to these sentences, is segmented, part-of-speech tagging and semantic analysis.To the text data of input according to punctuate
Symbol carries out punctuate processing to text data, and each sentence is mainly divided into clause using Hanlp, then carries out to sentence semantic
Analysis.
Step 3: building major event emotion reason corpus；Emotion word is identified first, has mainly used Taiwan
University NTUSD simplified form of Chinese Character feeling polarities dictionary, it includes 11086 emotion words which, which has altogether, including 8276 cathode
Property word and 2810 positive polarity words.Then emotion reason is identified, mark granularity is clause's rank, uses punctuation mark
Sentence is divided into clause one by one, the mark of emotion reason is exemplified below, in example 1, the emotion reason of " pained " be " for
812 firemans sacrificed ".
Example 1: the fireman sacrificed for 812 is really very pained, it is desirable to the available compensation of their family members, motherland
It rises with force and spirit.
Step 4: constructing emotion reason ontology model according to ABC ontology model, mainly modeled according to event, thing
The content of part mainly passes through these conceptual descriptions of agent, action, situation, event, place, and agent is to cause
The entity that major event occurs, action are the movement in major event, and event is major event, and place is major event
Spot.
Step 5: the relationship by emotional event, region, behavior in fusion emotion reason field etc. between concepts and concept
Emotion reason ontology model is expanded, region, which is inherited, comes from ABC:place.Emotional event, which is inherited, comes from ABC:event.It represents
Major event on network social intercourse media.
Step 6: constructing emotion reason tree according to ternary relation；When mood reason event structure defines, event clear first
The argument that event trigger word is directed toward then is acted both implementer and action object and is included in mood original by the importance of trigger word
Because obtaining elementary event triple in event basic structure.From the angle of identification triple element, this semantic character labeling
(Semantic Role Labeling, SRL) task objective is very close.Semantic character labeling is a kind of Shallow Semantic Parsing skill
Art, it is therefore an objective to which analysis obtains the semanteme that predicate-argument is carried in sentence, such as actor, movement, word denoting the receiver of an action person, tool and place
Deng.
Step 7: judging whether read statement contains emotion word；Taiwan Univ. NTUSD simplified form of Chinese Character emotion is mainly used
Polarity dictionary, it includes 11086 emotion words which, which has altogether, including 8276 negative polarity words and 2810 positive polarity words
Language.
Step 8: being extracted using algorithm of support vector machine to emotion reason event, using embeding to input
Sentence carries out input processing, then as the input vector of support vector machines.Judge to contain when sentence kind using support vector machines
There is emotion reason.
Step 9: recall rate R value and F value assess emotional event extraction algorithm using accurate rate P value.
Step 10: analyzing on emotion reason ontology model emotion reason, walking for great public sentiment event is analyzed
To.
Embodiment 2
The proposed in this paper kind of emotion reason abstracting method based on ontology model and multi-kernel support vector machine, mainly finds
Emotion reason in text.Viewing people can be excavated to different film emotion reasons, advised for film market construction.
Step 1: using JAVA language under the environmental level of myeclipse, using Struts, Spring,
The framework technologies such as Hibernate carry out realizing a crawler frame, the film review news data of bean cotyledon film are crawled, as this method
Input data.The method that crawls mainly is divided into three steps, first according to kind of a sublink, extracts Object linking and is put into team to be crawled
Column；Then the information that needs are parsed and extracted from the page, finally handles data, as a file format by the data extracted
Storage or deposit database and search engine index library etc..
Step 2: being pre-processed to data, in the process, is then made pauses in reading unpunctuated ancient writings to these sentences, segmented, part of speech mark
Note and semantic analysis.Punctuate processing is carried out to text data according to punctuation mark to the text data of input, mainly utilizes Hanlp
Each sentence is divided into clause, semantic analysis then is carried out to sentence.
Step 3: building major event emotion reason corpus；Emotion word is identified first, has mainly used Taiwan
University NTUSD simplified form of Chinese Character feeling polarities dictionary, it includes 11086 emotion words which, which has altogether, including 8276 cathode
Property word and 2810 positive polarity words.Then emotion reason is identified, mark granularity is clause's rank, uses punctuation mark
Sentence is divided into clause one by one, the mark of emotion reason is exemplified below, and in example 2, the emotion reason of " liking " is " " one
Good play out " film of this somewhat hazy and illusionary color is really very good-looking ".
Example 2: the film of " one goes out good play " this somewhat hazy and illusionary color is really very good-looking, likes！Support Huang Bo director！
Step 4: constructing emotion reason ontology model according to ABC ontology model, mainly modeled according to event, thing
The content of part mainly passes through these conceptual descriptions of agent, action, situation, event, place, and agent is to cause
The entity that major event occurs, action are the movement in major event, and event is major event, and place is major event
Spot.
Step 5: the relationship by emotional event, region, behavior in fusion emotion reason field etc. between concepts and concept
Emotion reason ontology model is expanded, region, which is inherited, comes from ABC:place.
Step 6: constructing emotion reason tree according to ternary relation；When mood reason event structure defines, event clear first
The argument that event trigger word is directed toward then is acted both implementer and action object and is included in mood original by the importance of trigger word
Because obtaining elementary event triple in event basic structure.From the angle of identification triple element, this semantic character labeling
(Semantic Role Labeling, SRL) task objective is very close.Semantic character labeling is a kind of Shallow Semantic Parsing skill
Art, it is therefore an objective to which analysis obtains the semanteme that predicate-argument is carried in sentence, such as actor, movement, word denoting the receiver of an action person, tool and place
Deng.
Step 7: judging whether read statement contains emotion word；Taiwan Univ. NTUSD simplified form of Chinese Character emotion is mainly used
Polarity dictionary, it includes 11086 emotion words which, which has altogether, including 8276 negative polarity words and 2810 positive polarity words
Language.
Step 8: being extracted using algorithm of support vector machine to emotion reason event, using embeding to input
Sentence carries out input processing, then as the input vector of support vector machines.Judge to contain when sentence kind using support vector machines
There is emotion reason.
Step 9: recall rate R value and F value assess emotional event extraction algorithm using accurate rate P value.
Step 10: analyzing on emotion reason ontology model emotion reason.
Embodiment 3
The proposed in this paper kind of emotion reason abstracting method based on ontology model and multi-kernel support vector machine, mainly finds
Emotion reason in text.Construction of the tourist to the emotion and emotion reason at certain sight spots, to the sight spot of tourism can be excavated
It presents one's view.
Step 1: using JAVA language under the environmental level of myeclipse, using Struts, Spring,
The framework technologies such as Hibernate carry out realizing a crawler frame, tourist's comment data of way ox net are crawled, as this method
Input data.The method that crawls mainly is divided into three steps, first according to kind of a sublink, extracts Object linking and is put into queue to be crawled；
Then the information that needs are parsed and extracted from the page, finally handles data, and the data extracted are deposited as a file format
Put or be stored in database and search engine index library etc..
Step 2: being pre-processed to the text data of way ox net, is then made pauses in reading unpunctuated ancient writings to these sentences, segmented, part of speech
Mark and semantic analysis.Punctuate processing is carried out to text data according to punctuation mark to the text data of input, it is main to utilize
Each sentence is divided into clause by Hanlp, then carries out semantic analysis to sentence.
Step 3: building major event emotion reason corpus；Emotion word is identified first, has mainly used Taiwan
University NTUSD simplified form of Chinese Character feeling polarities dictionary, it includes 11086 emotion words which, which has altogether, including 8276 cathode
Property word and 2810 positive polarity words.Then emotion reason is identified, mark granularity is clause's rank, uses punctuation mark
Sentence is divided into clause one by one, the mark of emotion reason is exemplified below, and in example 3, the emotion reason of " unhappy " is " just
It is that the people that goes on May Day is too many ".
Example 3: Great Wall is pretty good object for appreciation, is exactly that the people that goes on May Day is too many, unhappy！
Step 4: constructing emotion reason ontology model according to ABC ontology model, mainly modeled according to event, thing
The content of part mainly passes through these conceptual descriptions of agent, action, situation, event, place, and agent is to cause
The entity that major event occurs, action are the movement in major event, and event is major event, and place is major event
Spot.
Step 5: the relationship by emotional event, region, behavior in fusion emotion reason field etc. between concepts and concept
Emotion reason ontology model is expanded, region, which is inherited, comes from ABC:place.Emotional event, which is inherited, comes from ABC:event.
Step 6: constructing emotion reason tree according to ternary relation；When mood reason event structure defines, event clear first
The argument that event trigger word is directed toward then is acted both implementer and action object and is included in mood original by the importance of trigger word
Because obtaining elementary event triple in event basic structure.From the angle of identification triple element, this semantic character labeling
(Semantic Role Labeling, SRL) task objective is very close.Semantic character labeling is a kind of Shallow Semantic Parsing skill
Art, it is therefore an objective to which analysis obtains the semanteme that predicate-argument is carried in sentence, such as actor, movement, word denoting the receiver of an action person, tool and place
Deng.
Step 7: judging whether read statement contains emotion word；Taiwan Univ. NTUSD simplified form of Chinese Character emotion is mainly used
Polarity dictionary, it includes 11086 emotion words which, which has altogether, including 8276 negative polarity words and 2810 positive polarity words
Language.
Step 8: being extracted using algorithm of support vector machine to emotion reason event, using embeding to input
Sentence carries out input processing, then as the input vector of support vector machines.Judge to contain when sentence kind using support vector machines
There is emotion reason.
Step 9: recall rate R value and F value assess emotional event extraction algorithm using accurate rate P value.
Step 10: analyzing on emotion reason ontology model emotion reason, common people's centering medical treatment can analyze out
The emotion and emotion reason of system, the construction of Chinese Medical treatment system present one's view.
Claims (8)
1. the emotion reason abstracting method based on ontology model and multi-kernel support vector machine, which is characterized in that this method mainly walks
Suddenly are as follows: building emotion reason corpus first；Then emotion reason ontology model is constructed according to ABC ontology model；In this model
It is upper that emotion reason event is identified based on algorithm of support vector machine.
2. the emotion reason abstracting method according to claim 1 based on ontology model and multi-kernel support vector machine, special
Sign is, the specific steps are as follows:
1) crawler code is write, news data is crawled, each news data content includes: model title, model content, model
Deliver time, news url；
2) data are pre-processed, removes stop words, segmented；
3) emotion reason corpus is constructed；
4) emotion reason ontology model is constructed according to ABC ontology model；
5) by emotional event, emotion reason, people, region, behavior in fusion emotion reason field etc. between concepts and concept
Relationship expands emotion reason ontology model；
6) emotion reason tree is constructed according to ternary relation；
7) judge whether read statement contains emotion word；
8) emotion reason event is extracted using algorithm of support vector machine；
9) accurate rate P value is used, recall rate R value and F value assess emotional event extraction algorithm；
10) emotion reason is analyzed on emotion reason ontology model.
3. the emotion reason abstracting method according to claim 2 based on ontology model and multi-kernel support vector machine, special
Sign is, the step 1) is mainly using JAVA language under the environmental level of myeclipse, using including Struts,
Spring, Hibernate framework technology are realized.
4. the emotion reason abstracting method according to claim 2 based on ontology model and multi-kernel support vector machine, special
Sign is, carries out data preprocessing operation with existing Chinese word segmentation tool in the step 2), and pre-treatment step includes removing
Stop words, punctuate, participle, part-of-speech tagging or semantic analysis.
5. the emotion reason abstracting method according to claim 4 based on ontology model and multi-kernel support vector machine, special
Sign is that the punctuate processing is mainly realized using existing segmenter.
6. the emotion reason abstracting method according to claim 5 based on ontology model and multi-kernel support vector machine, special
Sign is, the segmenter mainly include word segmenter, Ansj segmenter, Stamford participle, Lucene&Nutch segmenter,
Stamford segmenter or Lucene&Nutch segmenter, Hanlp, Jieba.
7. the emotion reason abstracting method according to claim 2 based on ontology model and multi-kernel support vector machine, special
Sign is that step 4) the emotion reason ontology model is mainly according to preceding because of-behavior-consequence ABC ontology model progress structure
Build, ABC ontology model is mainly modeled according to event, the content of event mainly pass through agent, action,
Situation, event, place description.
8. the emotion reason abstracting method according to claim 2 based on ontology model and multi-kernel support vector machine, special
Sign is, the step 5) is specifically: to causing event factor to be extended, be extended to region, be extended to emotion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811303734.1A CN109670045A (en) | 2018-11-02 | 2018-11-02 | Emotion reason abstracting method based on ontology model and multi-kernel support vector machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811303734.1A CN109670045A (en) | 2018-11-02 | 2018-11-02 | Emotion reason abstracting method based on ontology model and multi-kernel support vector machine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109670045A true CN109670045A (en) | 2019-04-23 |
Family
ID=66141726
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811303734.1A Pending CN109670045A (en) | 2018-11-02 | 2018-11-02 | Emotion reason abstracting method based on ontology model and multi-kernel support vector machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109670045A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111382565A (en) * | 2020-03-09 | 2020-07-07 | 南京理工大学 | Multi-label-based emotion-reason pair extraction method and system |
| CN113011192A (en) * | 2021-03-16 | 2021-06-22 | 广东工业大学 | Text emotional feature extraction method based on attention causal explanation |
| CN115203412A (en) * | 2022-07-06 | 2022-10-18 | 腾讯科技（深圳）有限公司 | Emotion viewpoint information analysis method and device, storage medium and electronic equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105630970A (en) * | 2015-12-24 | 2016-06-01 | 哈尔滨工业大学 | Social media data processing system and method |
| CN106547866A (en) * | 2016-10-24 | 2017-03-29 | 西安邮电大学 | A kind of fine granularity sensibility classification method based on the random co-occurrence network of emotion word |
| CN107851093A (en) * | 2015-06-30 | 2018-03-27 | 微软技术许可有限责任公司 | The text of free form is handled using semantic hierarchy structure |
| CN107908715A (en) * | 2017-11-10 | 2018-04-13 | 中国民航大学 | Microblog emotional polarity discriminating method based on Adaboost and grader Weighted Fusion |
-
2018
- 2018-11-02 CN CN201811303734.1A patent/CN109670045A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107851093A (en) * | 2015-06-30 | 2018-03-27 | 微软技术许可有限责任公司 | The text of free form is handled using semantic hierarchy structure |
| CN105630970A (en) * | 2015-12-24 | 2016-06-01 | 哈尔滨工业大学 | Social media data processing system and method |
| CN106547866A (en) * | 2016-10-24 | 2017-03-29 | 西安邮电大学 | A kind of fine granularity sensibility classification method based on the random co-occurrence network of emotion word |
| CN107908715A (en) * | 2017-11-10 | 2018-04-13 | 中国民航大学 | Microblog emotional polarity discriminating method based on Adaboost and grader Weighted Fusion |
Non-Patent Citations (1)
| Title |
|---|
| 王文俊等: ""应急事件Ontology语义模型及其应用"", 《计算机工程》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111382565A (en) * | 2020-03-09 | 2020-07-07 | 南京理工大学 | Multi-label-based emotion-reason pair extraction method and system |
| CN111382565B (en) * | 2020-03-09 | 2023-08-11 | 南京理工大学 | Emotion-reason pair extraction method and system based on multiple labels |
| CN113011192A (en) * | 2021-03-16 | 2021-06-22 | 广东工业大学 | Text emotional feature extraction method based on attention causal explanation |
| CN113011192B (en) * | 2021-03-16 | 2023-09-15 | 广东工业大学 | Text emotion feature extraction method based on attention causal interpretation |
| CN115203412A (en) * | 2022-07-06 | 2022-10-18 | 腾讯科技（深圳）有限公司 | Emotion viewpoint information analysis method and device, storage medium and electronic equipment |
| CN115203412B (en) * | 2022-07-06 | 2024-02-20 | 腾讯科技（深圳）有限公司 | Emotion viewpoint information analysis method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Arora et al. | Character level embedding with deep convolutional neural network for text normalization of unstructured data for Twitter sentiment analysis | |
| Du et al. | Self-training improves pre-training for natural language understanding | |
| Wang et al. | Combining Knowledge with Deep Convolutional Neural Networks for Short Text Classification. | |
| Qi et al. | Openhownet: An open sememe-based lexical knowledge base | |
| Demir et al. | Improving named entity recognition for morphologically rich languages using word embeddings | |
| CN109800310B (en) | Electric power operation and maintenance text analysis method based on structured expression | |
| CN113987104B (en) | Generating type event extraction method based on ontology guidance | |
| CN108628828A (en) | A kind of joint abstracting method of viewpoint and its holder based on from attention | |
| Alam et al. | Bidirectional LSTMs—CRFs networks for bangla POS tagging | |
| Mehmood et al. | A precisely xtreme-multi channel hybrid approach for roman urdu sentiment analysis | |
| Suman et al. | Why pay more? A simple and efficient named entity recognition system for tweets | |
| Dudhabaware et al. | Review on natural language processing tasks for text documents | |
| Yang et al. | Deep learning and its applications to natural language processing | |
| Xian et al. | Self-guiding multimodal LSTM—when we do not have a perfect training dataset for image captioning | |
| CN109670045A (en) | Emotion reason abstracting method based on ontology model and multi-kernel support vector machine | |
| Wang et al. | Mongolian named entity recognition with bidirectional recurrent neural networks | |
| CN107526721A (en) | A kind of disambiguation method and device to electric business product review vocabulary | |
| Raza et al. | A review on Urdu language parsing | |
| CN106897274B (en) | Cross-language comment replying method | |
| Piskorski et al. | Exploring the usefulness of cross-lingual information fusion for refining real-time news event extraction: A preliminary study | |
| Nooralahzadeh et al. | Part of speech tagging for french social media data | |
| Uban et al. | Studying the evolution of scientific topics and their relationships | |
| Zhang et al. | Chinese-English mixed text normalization | |
| Ni et al. | Masked siamese prompt tuning for few-shot natural language understanding | |
| Celikyilmaz et al. | An empirical investigation of word class-based features for natural language understanding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190423 |