CN105069021B - Chinese short text sensibility classification method based on field - Google Patents
Chinese short text sensibility classification method based on field Download PDFInfo
- Publication number
- CN105069021B CN105069021B CN201510415825.4A CN201510415825A CN105069021B CN 105069021 B CN105069021 B CN 105069021B CN 201510415825 A CN201510415825 A CN 201510415825A CN 105069021 B CN105069021 B CN 105069021B
- Authority
- CN
- China
- Prior art keywords
- word
- short text
- field
- emotion
- sentiment dictionary
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/374—Thesaurus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
Abstract
The invention discloses a kind of Chinese short text sensibility classification method based on field, including：Data prediction, i.e. sentence segmentation, participle, stop words filtering and field division are carried out to short text；Build the field sentiment dictionary of domain-oriented；Using above-mentioned sentiment dictionary, and using corpus as data set, the TF IDF weight calculations of the extraction and matching in emotion path, the extraction of candidate word and polarity discriminating and emotion word are carried out；Extract short text affective characteristics；Corpus is trained using random forests algorithm or the short text of unknown affective style is differentiated.Experiment shows that the present invention suggests plans and has very high accuracy rate.
Description
Technical field
The present invention relates to machine learning techniques field, more particularly to a kind of Chinese short text emotional semantic classification side based on field
Method.
Background technology
Internet is developed rapidly so that social networks and electric business shopping platform are able to be subject to user's more and more widely
Favor, such as facebook, push away spy, Sina weibo, bean cotyledon, the domestic and international network platform in Jingdone district and Taobao.The number in these network platforms
According to increasing while explosion type is presented, including evaluation to commodity, to the view of event around and to life interesting episode or mood swing
Record etc..Wherein, short text is the common important form of these data, and often with emotional color or subjective consciousness.It is right
Emotion in this short text data expressed by user is excavated, help to allow different user object preferably certainly select or
Service, such as provides more pertinent recommendation in selection to user, provides more effectively service when promoting product to electric business, to
Government or department of news media, which provide, reliably predicts or pushes potential focus incident etc..
Text emotion analysis is popular in natural language processing (Natural Language Processing, NLP) field
Research direction, obtained the widely studied analysis of scholar.The technology proposed has very much, but can be divided mainly into 2 kinds：One kind is
Method based on sentiment dictionary, another kind are the methods based on machine learning.Method based on sentiment dictionary (is divided with emotion word
To be positive and passive) Main Basiss that differentiate as emotion, i.e., the emotion contained according to emotion word come decision-making text.Based on machine
The method of study is to be classified using the grader according to training to the emotion of text.Two kinds of technical solutions are favourable
Disadvantage：The former algorithm is often relatively simple, and algorithm complex is relatively low, and without a large amount of label corpus；But there are sentiment dictionary
Easily omit, ambiguity or extreme, and the emotion difference of the emotion word generation to different scenes can not usually perceive.The latter's accuracy rate
It is often high compared with the former, but training affective characteristics grader needs substantial amounts of tape label corpus, and corpus will be chosen suitably.
The content of the invention
The technical problems to be solved by the invention are how efficiently to combine sentiment dictionary and machine learning to Chinese short essay
This emotion is classified automatically, to improve text automatic marking training effectiveness and final classification utensil is had high-accuracy.
In order to solve the above technical problem, the present invention provides a kind of Chinese short text emotional semantic classification side based on field
Method, including：
Data prediction, including sentence segmentation, participle, stop words filtering and field division are carried out to short text；
Build the field sentiment dictionary of different field；
The emotion value of short text is calculated using data after the field sentiment dictionary and pretreatment；
Extract the affective characteristics of short text；
Affective characteristics according to being extracted uses random forest to be trained for classification tool to corpus or to not knowing
The short text of sense type is differentiated.
Further, it is described that data prediction, including sentence segmentation, participle, stop words filtering and neck are carried out to short text
Domain divides, and specifically includes：
Short text is divided into multiple sentences using punctuation mark；
ICTCLAS is used to segment instrument by the multiple sentence cutting as independent word；
The word of cutting is filtered using vocabulary is disabled；
According to short text and context environmental, with reference to domain lexicon, short text fields is marked off.
Further, the field sentiment dictionary of the structure different field, specifically includes：
The emotion word unrelated with field is picked out from existing sentiment dictionary, and therefrom deletes ambiguity and the word being of little use
Language, forms basic sentiment dictionary；
Extract noun all in corpus and be ranked up by word frequency, and choose the higher noun of word frequency using threshold method
As evaluation object；
Using between the modification emotion word in the dependency grammar analysis extraction evaluation object and the basic sentiment dictionary
All emotion paths；
According to all emotion paths, word corresponding with the emotion path that the evaluation object is consistent is matched, is being arranged
After the word in basic sentiment dictionary, vocabulary of the part of speech for adjective, adverbial word and verb will be obtained as candidate's emotion word；
After carrying out feeling polarities classification to candidate's emotion word using word similitude distinguished number, folded with basic dictionary
Add, form field sentiment dictionary.
Further, the emotion value of short text, specific bag are calculated using data after the field sentiment dictionary and pretreatment
Include：
The TF-IDF values of each word in the field sentiment dictionary are calculated, wherein, TF-IDF=TF*IDF, in formula,
TF represents word frequency, and IDF represents reverse document-frequency；
For the multiple words obtained after short text word segmentation processing, the emotion value of each word is calculated, i.e., according to word
TF-IDF values assign word different weights；
The weighted sum of the emotion value of all words is calculated, obtains the emotion value of short text.
Further, the emotion value for being directed to the multiple words obtained after short text word segmentation processing, calculating each word,
Different weights are assigned to word according to the TF-IDF values of word, are specifically included：
For the multiple words obtained after short text word segmentation processing, position and propensity value p that each word occurs are recorded, its
In, if word is positive, p initialization values are f (TF-IDF), if word is passiveness, p initialization values are-f (TF-IDF),
Wherein, f (TF-IDF) is the default initial emotion value of word；
The position occurred according to word, judges whether negative word occur between word, if occurring, calculates of negative word
Number, when the number of negative word be odd number, just by the propensity value p of the word behind negative word reversion, otherwise propensity value p is not
Become, final propensity value p is the emotion value of word；
Different weights are assigned to different words according to the TF-IDF values of word.
Further, it is described to use random forest to be instructed for classification tool to corpus according to the affective characteristics extracted
Practice or the short text of unknown affective style is differentiated, specifically include：
Using arrf feature templates by affective characteristics document formatting；
Random forests algorithm in weka is called to be trained as classification tool according to the affective characteristics of extracted corpus
Or emotion prediction classification is carried out to the short text of unknown affective style.
Implement the present invention, have the advantages that：
1) the short text emotion method of discrimination proposed by the present invention based on field improves the standard of text data emotional semantic classification
True rate；
2) propose that the accuracy rate that the sentiment dictionary based on field obtains can be reached apparently higher than using basic sentiment dictionary
The accuracy rate arrived.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is attached drawing needed in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
Fig. 1 is that the flow of one embodiment of the Chinese short text sensibility classification method provided by the invention based on field is shown
It is intended to；
Fig. 2 is the flow diagram of the specific steps of step S101 in Fig. 1；
Fig. 3 is the contrast and experiment figure of sentiment dictionary and traditional sentiment dictionary in method proposed by the invention.
Fig. 4 is the test result exemplary plot in four fields.
Embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without making creative work
Embodiment, belongs to the scope of protection of the invention.
Fig. 1 is that the flow of one embodiment of the Chinese short text sensibility classification method provided by the invention based on field is shown
It is intended to, includes the following steps：
S101, carry out short text data prediction, including sentence segmentation, participle, stop words filtering and field division.
Specifically, as shown in Fig. 2, step S101 includes step：
S1011, using punctuation mark be divided into multiple sentences by short text；
S1012, use ICTCLAS to segment instrument by the multiple sentence cutting as independent word；
S1013, using disable vocabulary the word of cutting is filtered；
S1014, according to short text and context environmental, with reference to domain lexicon, mark off short text fields.
S102, the field sentiment dictionary for building different field.
Specifically, step S102 includes step：
S1021, pick out the emotion word unrelated with field from existing sentiment dictionary, and has therefrom deleted ambiguity and seldom
Word, forms basic sentiment dictionary；
S1022, extract all noun in corpus and be ranked up by word frequency, and it is higher using threshold method to choose word frequency
Noun as evaluation object.
S1023, analyze the modification emotion extracted in the evaluation object and the basic sentiment dictionary using dependency grammar
All emotion paths between word；
S1024, according to all emotion paths, match word corresponding with the emotion path that the evaluation object is consistent
Language, after the word in excluding basic sentiment dictionary, will obtain vocabulary of the part of speech for adjective, adverbial word and verb as candidate's feelings
Feel word；
S1025, using word similitude distinguished number to candidate's emotion word carry out feeling polarities classification after, with basis
Dictionary is superimposed, and forms field sentiment dictionary.
S103, the emotion value using data calculating short text after the field sentiment dictionary and pretreatment.
Specifically, step S103 includes step：
S1031, the TF-IDF values for calculating each word in the field sentiment dictionary, wherein, TF-IDF=TF*IDF,
In formula, TF represents word frequency, and IDF represents reverse document-frequency；
S1032, for the multiple words obtained after short text word segmentation processing, calculate the emotion value of each word, i.e. basis
The TF-IDF values of word assign word different weights.
Specifically, step S1032 includes：
For the multiple words obtained after short text word segmentation processing, position and propensity value p that each word occurs are recorded, its
In, if word is positive, p initialization values are f (TF-IDF), if word is passiveness, p initialization values are-f (TF-IDF),
Wherein, f (TF-IDF) is the default initial emotion value of word；
The position occurred according to word, judges whether negative word occur between word, if occurring, calculates of negative word
Number, when the number of negative word be odd number, just by the propensity value p of the word behind negative word reversion, otherwise propensity value p is not
Become, final propensity value p is the emotion value of word；
Different weights are assigned to different words according to the TF-IDF values of word.
S1033, calculate all words emotion value weighted sum, obtain the emotion value of short text.
S104, the affective characteristics for extracting short text.
Wherein, affective characteristics specifically includes 9 features, as shown in table 1.
Table 1
S105, according to the affective characteristics extracted use random forest to be trained for classification tool to corpus or to not
Know that the short text of affective style is differentiated.
Specifically, step S105 includes step：
S1051, using arrf feature templates formatted affective characteristics；
S1052, call random forest in weka to be trained for classification tool to corpus or to unknown affective style
Short text is differentiated.
The embodiment of the present invention is emulated, obtain accuracy rate compared with the algorithm of Tan et al. as shown in table 2, in hotel
Field and books field, the present invention carry algorithm of the algorithm than Tan et al. and are improved in terms of accuracy rate very much, but are led in electronics
In domain, this research institute puies forward the accuracy rate of algorithm somewhat almost.
Table 2
Fig. 3 is the contrast and experiment of sentiment dictionary and basic sentiment dictionary in method proposed by the invention.The result shows that
Field sentiment dictionary is substantially better than the classifying quality of basic sentiment dictionary, and four field Average Accuracies improve 5.3%, wherein
4%, 5.2%, 2.9% and 8.8% has been respectively increased on books, hotel, electronic product and cinematic data collection.
Fig. 4 is the test result exemplary plot in four fields, and the wherein transverse axis of figure represents training set proportion, and the longitudinal axis is point
Class accuracy rate and F-Measure.By result it can be shown that classifying when training data is 80% and test data is 20%
Accuracy rate and F-Measure are best.
It should be noted that herein, term " comprising ", "comprising" or its any other variant are intended to non-row
His property includes, so that process, method, article or device including a series of elements not only include those key elements, and
And other elements that are not explicitly listed are further included, or further include as this process, method, article or device institute inherently
Key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that including this
Also there are other identical element in the process of key element, method, article or device.
The embodiments of the present invention are for illustration only, do not represent the quality of embodiment.
In several embodiments provided herein, the system and method illustrated can be real by another way
It is existing.For example, system embodiment described above is schematical；The division of the unit, is only that a kind of logic function is drawn
Point, there can be other dividing mode when actually realizing；Multiple units or component can combine or be desirably integrated into another
System, or some features can be ignored, or not perform.
Can directly it be held with reference to the step of method or algorithm that the embodiments described herein describes with hardware, processor
Capable software module, or the two combination are implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or use the present invention.
A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the scope of the invention.Therefore, the present invention will not
Be intended to be limited to the embodiments shown herein, and be to fit to it is consistent with the principles and novel features disclosed herein most
Wide scope.
Claims (5)
- A kind of 1. Chinese short text sensibility classification method based on field, it is characterised in that including：Data prediction, including sentence segmentation, participle, stop words filtering and field division are carried out to short text；Build the field sentiment dictionary of different field；The emotion value of short text is calculated using data after the field sentiment dictionary and pretreatment；Extract the affective characteristics of short text；Affective characteristics according to being extracted uses random forest to be trained for classification tool to corpus or to unknown emotion class The short text of type is differentiated；The field sentiment dictionary of the structure different field, specifically includes：The emotion word unrelated with field is picked out from existing sentiment dictionary, and therefrom deletes ambiguity and the word being of little use, The basic sentiment dictionary of composition；Extract all noun in corpus and be ranked up by word frequency, and by the use of threshold method choose the higher noun of word frequency as Evaluation object；The institute between the modification emotion word in the evaluation object and the basic sentiment dictionary is extracted using dependency grammar analysis There is emotion path；According to all emotion paths, word corresponding with the emotion path that the evaluation object is consistent is matched, is excluding base After word in plinth sentiment dictionary, vocabulary of the part of speech for adjective, adverbial word and verb will be obtained as candidate's emotion word；After carrying out feeling polarities classification to candidate's emotion word using word similitude distinguished number, it is superimposed with basic dictionary, Composition field sentiment dictionary.
- 2. the Chinese short text sensibility classification method based on field as claimed in claim 1, it is characterised in that described to short essay This progress data prediction, including sentence segmentation, participle, stop words filtering and field division, specifically include：Short text is divided into multiple sentences using punctuation mark；ICTCLAS is used to segment instrument by the multiple sentence cutting as independent word；The word of cutting is filtered using vocabulary is disabled；According to short text and context environmental, with reference to domain lexicon, short text fields is marked off.
- 3. the Chinese short text sensibility classification method based on field as claimed in claim 1, it is characterised in that utilize the neck Data calculate the emotion value of short text after domain sentiment dictionary and pretreatment, specifically include：The TF-IDF values of each word in the field sentiment dictionary are calculated, wherein, TF-IDF=TF*IDF, in formula, TF tables Show word frequency, IDF represents reverse document-frequency；For the multiple words obtained after short text word segmentation processing, the emotion value of each word is calculated, i.e., according to the TF- of word IDF values assign word different weights；The weighted sum of the emotion value of all words is calculated, obtains the emotion value of short text.
- 4. the Chinese short text sensibility classification method based on field as claimed in claim 3, it is characterised in that described for short The multiple words obtained after text word segmentation processing, calculate the emotion value of each word, i.e., according to the TF-IDF values of word to word Different weights are assigned, are specifically included：For the multiple words obtained after short text word segmentation processing, position and propensity value p that each word occurs are recorded, wherein, If word is positive, p initialization values are f (TF-IDF), if word is passiveness, p initialization values are-f (TF-IDF), its In, f (TF-IDF) is the default initial emotion value of word；The position occurred according to word, judges whether negative word occur between word, if occurring, calculates the number of negative word, When the number of negative word is odd number, just the propensity value p of the word behind negative word is inverted, otherwise propensity value p is constant, Final propensity value p is the emotion value of word；Different weights are assigned to different words according to the TF-IDF values of word.
- 5. the Chinese short text sensibility classification method based on field as claimed in claim 1, it is characterised in that described according to institute The affective characteristics of extraction uses random forest to be trained for classification tool to corpus or the short text to unknown affective style Differentiated, specifically included：Using arrf feature templates by affective characteristics document formatting；Random forests algorithm in weka is called to be trained as classification tool according to the affective characteristics of extracted corpus or right The short text of unknown affective style carries out emotion prediction classification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510415825.4A CN105069021B (en) | 2015-07-15 | 2015-07-15 | Chinese short text sensibility classification method based on field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510415825.4A CN105069021B (en) | 2015-07-15 | 2015-07-15 | Chinese short text sensibility classification method based on field |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105069021A CN105069021A (en) | 2015-11-18 |
| CN105069021B true CN105069021B (en) | 2018-04-20 |
Family
ID=54498394
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510415825.4A Expired - Fee Related CN105069021B (en) | 2015-07-15 | 2015-07-15 | Chinese short text sensibility classification method based on field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105069021B (en) |
Families Citing this family (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170213138A1 (en) * | 2016-01-27 | 2017-07-27 | Machine Zone, Inc. | Determining user sentiment in chat data |
| CN105808529B (en) * | 2016-03-10 | 2018-06-08 | 语联网（武汉）信息技术有限公司 | The method and apparatus that a kind of language material divides field |
| CN105930359A (en) * | 2016-04-11 | 2016-09-07 | 百度在线网络技术（北京）有限公司 | Tendency monitoring method and device |
| CN106096664B (en) * | 2016-06-23 | 2019-09-20 | 广州云数信息科技有限公司 | A kind of sentiment analysis method based on social network data |
| CN106202200B (en) * | 2016-06-28 | 2019-09-27 | 昆明理工大学 | A kind of emotion tendentiousness of text classification method based on fixed theme |
| CN108062300A (en) * | 2016-11-08 | 2018-05-22 | 中移(苏州)软件技术有限公司 | A kind of method and device that Sentiment orientation analysis is carried out based on Chinese text |
| CN106776686A (en) * | 2016-11-09 | 2017-05-31 | 武汉泰迪智慧科技有限公司 | Chinese domain short text understanding method and system based on many necks |
| CN106844516A (en) * | 2016-12-28 | 2017-06-13 | 中央民族大学 | A kind of extracting method and system of focus word |
| CN107194739B (en) * | 2017-05-25 | 2018-10-26 | 广州百奕信息科技有限公司 | A kind of intelligent recommendation system based on big data |
| CN107301171B (en) * | 2017-08-18 | 2020-09-01 | 武汉红茶数据技术有限公司 | Text emotion analysis method and system based on emotion dictionary learning |
| CN108399158B (en) * | 2018-02-05 | 2021-05-14 | 华南理工大学 | Attribute emotion classification method based on dependency tree and attention mechanism |
| CN108416375B (en) * | 2018-02-13 | 2020-07-07 | 中国联合网络通信集团有限公司 | Work order classification method and device |
| CN108874937B (en) * | 2018-05-31 | 2022-05-20 | 南通大学 | Emotion classification method based on part of speech combination and feature selection |
| CN109241276B (en) * | 2018-07-11 | 2022-03-08 | 河海大学 | Word classification method in text, and speech creativity evaluation method and system |
| CN109190106B (en) * | 2018-07-16 | 2023-01-10 | 中国传媒大学 | Emotional dictionary construction system and construction method |
| CN109214008A (en) * | 2018-09-28 | 2019-01-15 | 珠海中科先进技术研究院有限公司 | A kind of sentiment analysis method and system based on keyword extraction |
| CN109977396A (en) * | 2019-02-18 | 2019-07-05 | 深圳壹账通智能科技有限公司 | Emotion identification method, device, computer equipment and the computer storage medium of corpus participle |
| CN109840281A (en) * | 2019-02-27 | 2019-06-04 | 浪潮软件集团有限公司 | A kind of self study intelligent decision method based on random forests algorithm |
| CN110688836A (en) * | 2019-09-30 | 2020-01-14 | 湖南大学 | Automatic domain dictionary construction method based on supervised learning |
| CN111753525B (en) * | 2020-05-21 | 2023-11-10 | 浙江口碑网络技术有限公司 | Text classification method, device and equipment |
| CN112182332A (en) * | 2020-09-25 | 2021-01-05 | 科大国创云网科技有限公司 | Emotion classification method and system based on crawler collection |
| CN112347259A (en) * | 2020-11-17 | 2021-02-09 | 河北工程大学 | Comment text sentiment analysis method combining dictionary and machine learning |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102760153A (en) * | 2011-04-21 | 2012-10-31 | 帕洛阿尔托研究中心公司 | Incorporating lexicon knowledge to improve sentiment classification |
| CN103365867A (en) * | 2012-03-29 | 2013-10-23 | 腾讯科技（深圳）有限公司 | Method and device for emotion analysis of user evaluation |
| CN103678278A (en) * | 2013-12-16 | 2014-03-26 | 中国科学院计算机网络信息中心 | Chinese text emotion recognition method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080249764A1 (en) * | 2007-03-01 | 2008-10-09 | Microsoft Corporation | Smart Sentiment Classifier for Product Reviews |
-
2015
- 2015-07-15 CN CN201510415825.4A patent/CN105069021B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102760153A (en) * | 2011-04-21 | 2012-10-31 | 帕洛阿尔托研究中心公司 | Incorporating lexicon knowledge to improve sentiment classification |
| CN103365867A (en) * | 2012-03-29 | 2013-10-23 | 腾讯科技（深圳）有限公司 | Method and device for emotion analysis of user evaluation |
| CN103678278A (en) * | 2013-12-16 | 2014-03-26 | 中国科学院计算机网络信息中心 | Chinese text emotion recognition method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105069021A (en) | 2015-11-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105069021B (en) | Chinese short text sensibility classification method based on field | |
| Kar et al. | No rumours please! a multi-indic-lingual approach for covid fake-tweet detection | |
| Davidov et al. | Enhanced sentiment learning using twitter hashtags and smileys | |
| CN106407236B (en) | A kind of emotion tendency detection method towards comment data | |
| CN105786991B (en) | In conjunction with the Chinese emotion new word identification method and system of user feeling expression way | |
| CN104008091B (en) | A kind of network text sentiment analysis method based on emotion value | |
| CN104778209B (en) | A kind of opining mining method for millions scale news analysis | |
| Luyckx | Scalability issues in authorship attribution | |
| Llewellyn et al. | Summarizing newspaper comments | |
| CN106547789B (en) | Lyric generation method and device | |
| CN104298665A (en) | Identification method and device of evaluation objects of Chinese texts | |
| Brahimi et al. | Data and Text Mining Techniques for Classifying Arabic Tweet Polarity. | |
| KR101713558B1 (en) | Method of classification and analysis of sentiment in social network service | |
| CN103995853A (en) | Multi-language emotional data processing and classifying method and system based on key sentences | |
| KR101491628B1 (en) | Method, apparatus and system for extracting keyword affecting for mood change of the public using blog | |
| Abdelali et al. | QADI: Arabic dialect identification in the wild | |
| CN107203520A (en) | The method for building up of hotel's sentiment dictionary, the sentiment analysis method and system of comment | |
| Yüksel et al. | Turkish tweet classification with transformer encoder | |
| CN109657064A (en) | A kind of file classification method and device | |
| CN109325223A (en) | Article recommended method, device and electronic equipment | |
| Elouardighi et al. | A machine Learning approach for sentiment analysis in the standard or dialectal Arabic Facebook comments | |
| CN110119443A (en) | A kind of sentiment analysis method towards recommendation service | |
| CN109284381A (en) | The aspect viewpoint of fusion emoticon library and topic model passes judgement on attitude method for digging | |
| Hasan et al. | Sentiment classification in bangla textual content: A comparative study | |
| Tembhurnikar et al. | Topic detection using BNgram method and sentiment analysis on twitter dataset |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180420 Termination date: 20190715 |