CN103995803A - Fine granularity text sentiment analysis method - Google Patents
Fine granularity text sentiment analysis method Download PDFInfo
- Publication number
- CN103995803A CN103995803A CN201410178056.6A CN201410178056A CN103995803A CN 103995803 A CN103995803 A CN 103995803A CN 201410178056 A CN201410178056 A CN 201410178056A CN 103995803 A CN103995803 A CN 103995803A
- Authority
- CN
- China
- Prior art keywords
- emotion
- sentence
- word
- fine granularity
- benchmark
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The invention discloses a fine granularity text sentiment analysis method which includes the first step of building a fine granularity sentiment dictionary, the second step of judging a sentence structure relation and the third step of carrying out sentiment value assessment of a simple sentence. Sentiment related information of more users included by a text can be extracted, inner feelings of the users can be depicted better, and the method can be used for supporting related application researches such as user mood state and change condition analysis based on health.
Description
Technical field
The invention belongs to English text sentiment analysis technical field, relate to a kind of fine granularity text emotion analytical approach, specifically, relate to a kind of fine granularity sentiment analysis method for comment text.
Background technology
The mankind's emotion is complicated and many-sided.Due to the complicacy of emotion and with the relation of other something outsides, it belongs to the challenging phenomenon of tool in psychology.The traditional approach of current mood of understanding a people has multiple: as seeked advice from its subjective feeling, observe the variation in its countenance or behavior, with and physiological change.In fact, a people's mood is complicated, can not directly be measured, and only can identify by their external expressive form, so just expedites the emergence of out various for identifying human emotion's method.In the ordinary course of things, the method for a people's of modal identification emotional reactions roughly can be divided three classes: (1) self-report, (2) physiology method, (3) behavior observation.
Monitoring individual emotional state based on classic method needs a large amount of man power and materials, is difficult to obtain the long mood related data of a large number of users.Along with the development of online social networks, its number of users having constantly increases, people start custom and often the finding of oneself are felt and is shared with good friend on line, researcher can obtain a large number of users mood related data by the API of social network sites thus, extracts user's emotional state based on text emotion analytical technology.
Text emotion analysis is an emerging research topic, has very large researching value and using value.Patent 200910219161.9 is estimated topic language model according to the Expression of language of different themes text, calculates the distance of language model and the positive and negative emotion model of pending text, and the emotion of the nearest emotion model of selected distance is inclined to and is given the text.Patent 200910083522.1 determines that according to the label of training text the initial emotion of test text divides, utilize the emotion of test text described in the initial emotion point iterative computation of described test text divide and be normalized based on figure sort algorithm, to solve cross-cutting emotion tendentiousness of text problem analysis.Patent 201210088366.X judges the polarity of all sentences that comprise descriptor based on positive and negative sentiment dictionary, front sentence polarity sum and negative sentence polarity sum in result of calculation set, thus draw the emotion tendency of whole piece microblogging.Patent 201310000734.5 by construct one have Two-Level (bilayer) structure DCRF model realization the emotion tendency judgement of entity level, patent 201310036034.1 is utilized statistics and the calculating of the relational implementation fine granularity emotion intensity quantification between related information and emotion word and the qualifier between object properties and emotion word.
The user feeling that current existing sentiment analysis technology mainly contains text packets is divided into two classes: forward and negative sense, in the text emotion analysis that belongs to coarseness aspect the division of emotion classification, lose a large number of users emotion relevant information.In order fully to obtain the contained information of user comment, better portray user's impression, the present invention does further fine granularity sentiment analysis to comment text, and by positive negative tendency Further Division respectively, for example negative emotions can be angry, can be also sadness etc.
Summary of the invention
The object of the invention is to overcome the defect that above-mentioned technology exists, a kind of fine granularity text emotion analytical approach is provided, the method can more fully obtain the contained emotion information of user comment, can better support relevant applied research, for example the user emotion state based on healthy and situation of change analysis.
Its concrete technical scheme is:
Step 1: build fine granularity sentiment dictionary
Choose international generally acknowledged benchmark emotional semantic classification as fine granularity emotional semantic classification, and using benchmark emotion word as seed emotion word of all categories, by wordNet (by the psychologist of Princeton university, a kind of English dictionary based on cognitive linguistics of linguist and Computer Engineer's co-design) search its synonym set, and put into corresponding classification, complete the first step enlarging of fine granularity sentiment dictionary;
Word is divided into four classes by wordNet: noun, verb, adverbial word and adjective; Expanded the noun disposition sense set obtaining by benchmark emotion word, and according to identical mode according to the adjective of benchmark emotion word, verb and adverbial word form, be built into respectively the emotion set of its adjective, verb and adverbial word form; Generic emotion set, except the difference of part of speech, does not affect the calculating of emotion value, the emotion set under a classification is considered as to a large class, thereby completes the second step enlarging of fine granularity sentiment dictionary;
So far, the fine granularity sentiment dictionary of structure also cannot cover most emotion vocabulary; How all the other emotion words are referred to the problem of fine granularity emotion classification, are converted to and analyze itself and benchmark emotion word similarity based on general knowledge on concept hierarchy, and be assigned in the emotion classification of the benchmark emotion word representative that similarity is the highest; Result is sorted out in ultimate analysis, and improves the defect that may exist; So far complete the enlarging of fine granularity sentiment dictionary;
Step 2: sentence structure relation judgement
Judge in statement whether have conjunction, if had, represent that this sentence is compound sentence, obtain sentence structure relation that this conjunction represents and the computation rule of statement emotion value according to relation rule between sentence; If no, this statement is simple sentence;
Step 3: the emotion value evaluation of simple sentence
If compound sentence is split as two subordinate sentences and processes; Simple sentence if, directly calculates its emotion value; Now, consider the emotion value assessment method of simple sentence, comment emotion is calculated and will be considered descriptor correlativity, and the emotion word irrelevant with descriptor can calculate and bring interference to emotion; And theme mainly embodies by the theme (subject and object) of statement, only need name part of speech and the adjective affective characteristic words of consideration and Topic relative; According to sentence structure, dependence, emotion word degree of passing judgement on and relevant adverbs modify intensity, calculate the emotion value of simple sentence;
The imperfect short sentence often occurring for comment text, utilizes word part of speech, improves the accuracy of dependence judgement; In the time that the adjunctival before emotion word is long, sentence structure, word part of speech and dependence are combined, specific algorithm is as follows: first survey dependence, find out descriptor, and then find out the modified relationship that depends on descriptor, obtain the nominal phrase of descriptor and its adjunctival formation according to the result of sentence structure analysis, then analyze the structure of this nominal phrase and the part of speech of the word that adjunctival comprises, draw correct modified relationship;
Step 4: comment text fine granularity emotion is calculated
Obtain statement emotion value in conjunction with relation between sentence pattern and sentence; The overall emotion value that the emotion sum of all statements is comment text.
Compared with prior art, the invention has the beneficial effects as follows: can extract the more user feeling relevant information that text comprises, can better portray the impression of user's heart, for supporting relevant applied research, for example the user emotion state based on healthy and situation of change analysis.
Brief description of the drawings
Fig. 1 is fine granularity sentiment dictionary construction method process flow diagram of the present invention;
Fig. 2 is text fine granularity sentiment analysis method flow diagram of the present invention;
Fig. 3 is the example sentence sentence structure figure in example of the present invention.
Embodiment
Below in conjunction with the drawings and specific embodiments, technical scheme of the present invention is described in more detail.
As shown in Figure 1, detailed step is as follows for fine granularity sentiment dictionary construction method realization flow of the present invention:
Step 101: set benchmark emotion classification and seed emotion word.
Up to now, neither one the recognized standard is gone back in the division of psychology bound pair emotion, the present invention is taking the famous 6 benchmark emotions of scholar Ekman as example, specifically comprise: happiness (happiness), sadness (sadness), anger (anger), fear (fear), surprise (pleasantly surprised) and disgust (detest).First according to using 6 benchmark emotion words as seed emotion word of all categories, search its synonym set by wordNet, and put into corresponding classification, complete the first step enlarging of fine granularity sentiment dictionary.
Step 102: according to synonym expanding sentiment dictionary.
Word is divided into four classes by wordNet: noun, verb, adverbial word and adjective.We have obtained expanding by 6 benchmark emotion words the noun disposition sense set obtaining, and the mode according to identical, according to the adjective of 6 benchmark emotion words, verb and adverbial word form, is built into respectively to the emotion set of its adjective, verb and adverbial word form.For example, " joy (joy) " and " joyful (happy) " all belongs to " happiness (happiness) " emotion classification, but belong to two emotion set, joy belongs to a part of speech set, ioyful belongs to and describes part of speech set.Generic emotion set, except the difference of part of speech, does not affect the calculating of emotion value, and the emotion set under a classification is considered as a large class by the present invention, thereby completes the second step enlarging of fine granularity sentiment dictionary.
Step 103: based on general general knowledge storehouse expanding sentiment dictionary.
So far fine granularity sentiment dictionary comprises more than 1000 word altogether, and this is obviously not enough for analyzing text emotion tendency, also has the word of a large amount of expression people emotions not capped.For example give expression to the verb cry (crying) of obvious emotion, it cannot, by the mode of emotion set seed word synonym expansion above, join sentiment dictionary.Based on general knowledge, we can think the associated larger of cry (crying) and sad (sadness) and angry (anger) conventionally, and it often gives expression to the sad or angry mood of main body.When carry out the classification of fine granularity emotion word for cry, we are converted into emotion similarity based on general knowledge on concept hierarchy of analyzing cry (crying) and six emotional semantic classification representatives, similarity is higher, it is larger that we think that main body wants to give expression to the probability of such emotion, cry (crying) joined in the sentiment dictionary of such emotion.Calculate emotion word and emotional semantic classification similarity based on commonsense reasoning on concept hierarchy, general general knowledge storehouse is best selection, and the common sense knowledge of reflection emotion is a subset of general knowledge in these general knowledge storehouses.The present invention utilizes general general knowledge storehouse to calculate the similarity of emotion word and emotional semantic classification benchmark word, and is assigned in the emotion classification that similarity is the highest.
Step 104: set benchmark emotion classification and seed emotion word.
So far, the sentiment dictionary building can cover most emotion words, but according to the analysis to matrix of consequence, also there is obvious defect, we find the adjective for some, although the similarity result on concept hierarchy meets general knowledge substantially for its prototype and benchmark emotion word, its comparative degree and the highest similarity result obtaining are all but 0.For solving problems, we set up comparative degree and highest vocabulary for conventional adjective, and its comparative degree and the superlative degree are inherited classification and the similarity identical with prototype.So far complete the enlarging of fine granularity sentiment dictionary.
As shown in Figure 2, detailed step is as follows for text fine granularity sentiment analysis method realization flow of the present invention:
Step 201: the structure of fine granularity sentiment dictionary.
By setting benchmark emotion classification and seed emotion word, complete the structure of sentiment dictionary according to four parts such as synonym expanding sentiment dictionary, defect correction based on general general knowledge storehouse expanding sentiment dictionary, sentiment dictionary based on having set up, detailed step is referring to step 101～104.
Step 202: sentence structure relation judgement.
Judge in statement whether have conjunction, if had, represent that this sentence is compound sentence, obtain sentence structure relation that this conjunction represents and the computation rule of statement emotion value according to relation rule between sentence.If no, this statement is simple sentence.
In an English sentence, often there is the conjunction of expressing different relations, for example but (turnover), if (hypothesis), so (cause and effect), and (side by side), moreover (going forward one by one).Different conjunctions can produce different impacts to the emotion of sentence, and the present invention applies following rule and calculates:
(1) conjunction rule in sentence:
1) based on the conventional grammer custom of people, if there is no but, a tendency expressed in a general sentence.For comprising, table is arranged side by side, the sentence of the conjunction of progressive relationship, and the rule of applying in document is given different weights.
For example: this sentence of The camera takes great pictures and has a long battery life. means: this camera is taken a picture effective, and battery life is long.Conventionally,, if we know that great is forward, long is also generally forward so
(2) conjunction rule between sentence:
1), based on hypothesis, people also often express identical tendency between sentence.Unless there is but (still), however (but) etc.
2) for the conjunction that comprises different relations, we give its guiding clause different weights, the subordinate clause of progressive relationship conjunction guiding has the tendency that strengthens emotion tendency in general, the affectional variation of subordinate clause of cause and effect and the guiding of coordination conjunction is not very large, and hypothesis is related to the subordinate clause imagination to realistic situation often of conjunction guiding, its precondition has played prior effect in language performance, generally need to weaken the rear half point of hypothetical sentence.The setting of the weighing factor of the emotion tendency of the subordinate clause for other conjunction to its guiding, the present invention's application general rule.
3), for other universal relation conjunctions, as represented, precedence conjunction first (first) can not affect emotion tendency and the intensity of the subordinate clause of its guiding substantially; For the conjunction of table time, as when (when ... time) do not consider its emotion tendency.
The conjunction of table turning relation, taking but (still) as example, no matter as conjunction between conjunction in sentence or sentence, the subordinate clause contrast previous contents of its guiding has turnover semantic, but the positive and negative tendency of its emotion is but not necessarily contrary, while analyzing text emotion tendency, the emotion tendency of the subordinate clause of but (still) guiding, not necessarily contrary with the sentence emotion tendency before it, there are by analysis three kinds of situations:, subordinate clause consistent with the sentence tendency before it does not have obvious emotion tendency, contrary with the sentence tendency before it.Be exemplified below respectively:
1, the first situation: I ' m sure the wines are amazing too but I didn ' t regret getting a pint of the IPA.
Sentence justice: I determine that those wine are also very good, but I have not regretted selecting this beverage of a pint.
2, the second situation: Cool to see but don ' t get too close.
Sentence justice: this seems very cruel but does not lean on too closely.
3, the third kind of situations: It′s?not?that?cheap?as?they?say，But?good?for?who?are?crazy?with?brands！
Sentence justice: what they did not say is so cheap.But fine for liking ardently the people of brand.
Processing rule for above three kinds of situations:
If the subordinate clause of but (still) guiding, consistent with the short sentence tendency before it, its subordinate clause has the implication of going forward one by one, and increases the weight of the weight of its subordinate clause; If the subordinate clause of but (still) guiding does not have obvious emotion tendency, may be only as explanation or prompting etc., only consider the emotion tendency of its previous sentence; If the subordinate clause of but (still) guiding is contrary with the subordinate clause tendency before it, there is emotion reversion, but (still) the subordinate sentence meaning above can weaken, and gives prominence to the subordinate sentence after transferring.
Step 203: the emotion value evaluation of simple sentence.
In a sentence, the emotion of emotion phrase tendency and intensity are determined by the emotion word comprising and the ornamental equivalent that depends on thereon.The present invention utilizes Stanford Parser (parser of Stamford), extracts the dependence pair between the inner each participle of sentence unit.The main relation pair that the present invention is applied to, as shown in table 1.
The main dependence pair of table 1
| Title | Relation |
| nsubj | Subject and predicate |
| dobj | Object and predicate |
| amod | Adjective is modified |
| advmod | Adverbs modify |
| comod | Coordination |
| neg | Negate to modify |
This is noun, verb, adverbial word and adjective for base in sentiment dictionary, in English text, adverbial word is generally used for and modifies verb and adjective, emotion word only likely appears at noun, and in these three kinds of words of verb and adjective, inquiry sentiment dictionary can be found out its corresponding feeling polarities.For adverbial word, we build degree adverb table, include conventional adverbial word and corresponding modification intensity thereof.
Rule-based text emotion analysis, the emotion tendency of its sentence is that the emotion tendency by calculating emotion word in sentence gets, if but there is negative word before emotion word, can shine into impact to emotion tendency or the intensity of sentence, for example " I do not like it. (I do not like it) ", like (liking) itself expresses positive mood, but because there is not (no) to modify before it, there is reversal of poles in the emotion of whole sentence tendency, is transformed into negative mood.By analysis, uncertainty relation there will be on verb and adjective disposition sense word, may produce two kinds of impacts in degree: 1, reversal of poles occurs emotion tendency, and intensity is substantially unaffected on its emotion tendency.2, emotion tendency is constant, but remitted its fury.
Situation one: Try it, you won ' t regret.
Sentence justice: try, you can not regret.
Situation two: Not the best restaurant but always have something to eat.
Sentence justice: be not best restaurant, but still have thing to eat.
Comment on compared with general article statement, there is an outstanding feature to be: often to occur imperfect brief comment, be that sentence structure is imperfect, lacking one or several in subject, predicate or object, for example sentence ' Highly recommended. (strongly recommending) ' lacks subject and object.Incomplete sentence structure, strengthen the detection difficulty of dependence, cause surveying the decline of accuracy, taking previous sentence, ' Highly recommended. (strongly recommend) ' is example, because this does not have complete sentence structure, the correct dependence of its impalpable of syntactic analysis Highly (strongly) and recommended (recommendation), takeing for Highly is a noun, dependence mistake is judged to be to nsubj (subject-predicate relation), while causing calculating, only consider the emotion value of recommended, and do not consider the booster action of the emotion intensity of Highly to it.For this incomplete short sentence, the present invention takes first to obtain the part of speech of word, the method for recycling dependence.
In the time that the adjunctival before an emotion word is long, cause the difficulty of surveying dependence to strengthen, the accuracy of result of detection reduces.Taking sentence ' There is a completely unexpected great hotel. ' is example, the original idea of sentence is to say: fabulous (great) hotel that has complete (completely) (unexpected) beyond imagination here.Completely is used for modifying unexpected, Completely unexpected modifies great, completely unexpected great modifies hotel, because the level of modifying is darker, cause the detection of dependence to occur mistake, the detection of mistake is that completely unexpected modifies hotel, and great modifies separately hotel, has affected the accuracy that emotion intensity is calculated.But can see significantly that by this sentence structure (as shown in Figure 3) a completely unexpected great hotel is the nominal phrase based on descriptor hotel, and completely and unexpected form adverbial phrase modification great jointly.
Finally, according to the dependence detecting, extract the descriptor of statement, according to sentence structure, dependence, emotion word degree of passing judgement on and relevant adverbs modify intensity, calculate and depend on the emotion value of descriptor, thereby draw the emotion value of simple sentence.
Step 204: comment text fine granularity emotion is calculated
English is divided into Four types by purposes: declarative sentence, imperative sentence, exclamative sentence, interrogative sentence.Imperative sentence generally represents request, order etc., in comment text, seldom occurs, the present invention does not consider the emotion tendency of this sentence pattern.For usually occurring declarative sentence, exclamative sentence and interrogative sentence in comment, utilize punctuation mark and sentence pattern keyword to survey, and give different weights.
Obtain statement emotion value in conjunction with relation between sentence pattern and sentence.The overall emotion value that the emotion sum of all statements is comment text.
The above; it is only preferably embodiment of the present invention; protection scope of the present invention is not limited to this; any be familiar with those skilled in the art the present invention disclose technical scope in, the simple change of the technical scheme that can obtain apparently or equivalence replace all fall within the scope of protection of the present invention.
Claims (1)
1. a fine granularity text emotion analytical approach, is characterized in that, comprises the following steps:
Step 1: build fine granularity sentiment dictionary
Choose international generally acknowledged benchmark emotional semantic classification as fine granularity emotional semantic classification, and using benchmark emotion word as seed emotion word of all categories, search its synonym set by wordNet, and put into corresponding classification, complete the first step enlarging of fine granularity sentiment dictionary;
Word is divided into four classes by wordNet: noun, verb, adverbial word and adjective; Expanded the noun disposition sense set obtaining by benchmark emotion word, and according to identical mode according to the adjective of benchmark emotion word, verb and adverbial word form, be built into respectively the emotion set of its adjective, verb and adverbial word form; Generic emotion set, except the difference of part of speech, does not affect the calculating of emotion value, the emotion set under a classification is considered as to a large class, thereby completes the second step enlarging of fine granularity sentiment dictionary;
So far, the fine granularity sentiment dictionary of structure also cannot cover most emotion vocabulary; How all the other emotion words are referred to the problem of fine granularity emotion classification, are converted to and analyze itself and benchmark emotion word similarity based on general knowledge on concept hierarchy, and be assigned in the emotion classification of the benchmark emotion word representative that similarity is the highest; Result is sorted out in ultimate analysis, and improves the defect that may exist; So far complete the enlarging of fine granularity sentiment dictionary;
Step 2: sentence structure relation judgement
Judge in statement whether have conjunction, if had, represent that this sentence is compound sentence, obtain sentence structure relation that this conjunction represents and the computation rule of statement emotion value according to relation rule between sentence; If no, this statement is simple sentence;
Step 3: the emotion value evaluation of simple sentence
If compound sentence is split as two subordinate sentences and processes; Simple sentence if, directly calculates its emotion value; Now, consider the emotion value assessment method of simple sentence, comment emotion is calculated and will be considered descriptor correlativity, and the emotion word irrelevant with descriptor can calculate and bring interference to emotion; And theme mainly embodies by subject and the object of statement, only need name part of speech and adjective affective characteristic words that consideration is relevant to subject and object; According to sentence structure, dependence, emotion word degree of passing judgement on and relevant adverbs modify intensity, calculate the emotion value of simple sentence;
The imperfect short sentence often occurring for comment text, utilizes word part of speech, improves the accuracy of dependence judgement; In the time that the adjunctival before emotion word is long, sentence structure, word part of speech and dependence are combined, specific algorithm is as follows: first survey dependence, find out descriptor, and then find out the modified relationship that depends on descriptor, obtain the nominal phrase of descriptor and its adjunctival formation according to the result of sentence structure analysis, then analyze the structure of this nominal phrase and the part of speech of the word that adjunctival comprises, draw correct modified relationship;
Step 4: comment text fine granularity emotion is calculated
Obtain statement emotion value in conjunction with relation between sentence pattern and sentence; The overall emotion value that the emotion sum of all statements is comment text.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410178056.6A CN103995803B (en) | 2014-04-25 | 2014-04-25 | A kind of fine granularity text sentiment analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410178056.6A CN103995803B (en) | 2014-04-25 | 2014-04-25 | A kind of fine granularity text sentiment analysis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103995803A true CN103995803A (en) | 2014-08-20 |
| CN103995803B CN103995803B (en) | 2017-07-04 |
Family
ID=51309972
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410178056.6A Expired - Fee Related CN103995803B (en) | 2014-04-25 | 2014-04-25 | A kind of fine granularity text sentiment analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103995803B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881402A (en) * | 2015-06-02 | 2015-09-02 | 北京京东尚科信息技术有限公司 | Method and device for analyzing semantic orientation of Chinese network topic comment text |
| CN105335352A (en) * | 2015-11-30 | 2016-02-17 | 武汉大学 | Entity identification method based on Weibo emotion |
| CN107291689A (en) * | 2017-05-31 | 2017-10-24 | 温州市鹿城区中津先进科技研究院 | A kind of analysis method based on the Chinese network comments sentence theme semantic tendency of big data |
| CN107515856A (en) * | 2017-08-30 | 2017-12-26 | 哈尔滨工业大学 | A kind of fine granularity Emotion element abstracting method represented based on local message |
| CN107704996A (en) * | 2017-09-12 | 2018-02-16 | 青岛大学 | A kind of Teacher Evaluation System based on sentiment analysis |
| CN107729374A (en) * | 2017-09-13 | 2018-02-23 | 厦门快商通科技股份有限公司 | A kind of extending method of sentiment dictionary and text emotion recognition methods |
| CN107807920A (en) * | 2017-11-17 | 2018-03-16 | 新华网股份有限公司 | Construction method, device and the server of mood dictionary based on big data |
| CN107862087A (en) * | 2017-12-01 | 2018-03-30 | 广州简亦迅信息科技有限公司 | Sentiment analysis method, apparatus and storage medium based on big data and deep learning |
| CN108268439A (en) * | 2016-12-30 | 2018-07-10 | 北京国双科技有限公司 | The processing method and processing device of text emotion |
| CN108460015A (en) * | 2018-02-08 | 2018-08-28 | 合肥工业大学 | Text emotion grouped data enhances analysis method |
| CN109271623A (en) * | 2018-08-16 | 2019-01-25 | 龙马智芯（珠海横琴）科技有限公司 | Text emotion denoising method and system |
| CN110083684A (en) * | 2019-04-24 | 2019-08-02 | 吉林大学 | Interpretable recommended models towards fine granularity emotion |
| CN110765769A (en) * | 2019-08-27 | 2020-02-07 | 电子科技大学 | Entity attribute dependency emotion analysis method based on clause characteristics |
| CN112784011A (en) * | 2021-01-04 | 2021-05-11 | 南威软件股份有限公司 | Emotional problem processing method, device and medium based on CNN and LSTM |
| US11010564B2 (en) * | 2019-02-05 | 2021-05-18 | International Business Machines Corporation | Method for fine-grained affective states understanding and prediction |
| CN112905736A (en) * | 2021-01-27 | 2021-06-04 | 郑州轻工业大学 | Unsupervised text emotion analysis method based on quantum theory |
| CN115794988A (en) * | 2022-09-13 | 2023-03-14 | 广东美云智数科技有限公司 | Method, apparatus, and computer storage medium for extracting viewpoint of text |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
-
2014
- 2014-04-25 CN CN201410178056.6A patent/CN103995803B/en not_active Expired - Fee Related
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104881402B (en) * | 2015-06-02 | 2018-01-09 | 北京京东尚科信息技术有限公司 | The method and device of Chinese network topics comment text semantic tendency analysis |
| CN104881402A (en) * | 2015-06-02 | 2015-09-02 | 北京京东尚科信息技术有限公司 | Method and device for analyzing semantic orientation of Chinese network topic comment text |
| CN105335352A (en) * | 2015-11-30 | 2016-02-17 | 武汉大学 | Entity identification method based on Weibo emotion |
| CN108268439A (en) * | 2016-12-30 | 2018-07-10 | 北京国双科技有限公司 | The processing method and processing device of text emotion |
| CN108268439B (en) * | 2016-12-30 | 2021-09-07 | 北京国双科技有限公司 | Text emotion processing method and device |
| CN107291689A (en) * | 2017-05-31 | 2017-10-24 | 温州市鹿城区中津先进科技研究院 | A kind of analysis method based on the Chinese network comments sentence theme semantic tendency of big data |
| CN107515856A (en) * | 2017-08-30 | 2017-12-26 | 哈尔滨工业大学 | A kind of fine granularity Emotion element abstracting method represented based on local message |
| CN107704996A (en) * | 2017-09-12 | 2018-02-16 | 青岛大学 | A kind of Teacher Evaluation System based on sentiment analysis |
| CN107704996B (en) * | 2017-09-12 | 2021-07-02 | 青岛大学 | Teacher evaluation system based on emotion analysis |
| CN107729374A (en) * | 2017-09-13 | 2018-02-23 | 厦门快商通科技股份有限公司 | A kind of extending method of sentiment dictionary and text emotion recognition methods |
| CN107807920A (en) * | 2017-11-17 | 2018-03-16 | 新华网股份有限公司 | Construction method, device and the server of mood dictionary based on big data |
| CN107862087B (en) * | 2017-12-01 | 2022-02-18 | 深圳爱数云科技有限公司 | Emotion analysis method and device based on big data and deep learning and storage medium |
| CN107862087A (en) * | 2017-12-01 | 2018-03-30 | 广州简亦迅信息科技有限公司 | Sentiment analysis method, apparatus and storage medium based on big data and deep learning |
| CN108460015A (en) * | 2018-02-08 | 2018-08-28 | 合肥工业大学 | Text emotion grouped data enhances analysis method |
| CN109271623A (en) * | 2018-08-16 | 2019-01-25 | 龙马智芯（珠海横琴）科技有限公司 | Text emotion denoising method and system |
| US11010564B2 (en) * | 2019-02-05 | 2021-05-18 | International Business Machines Corporation | Method for fine-grained affective states understanding and prediction |
| US11132511B2 (en) | 2019-02-05 | 2021-09-28 | International Business Machines Corporation | System for fine-grained affective states understanding and prediction |
| CN110083684A (en) * | 2019-04-24 | 2019-08-02 | 吉林大学 | Interpretable recommended models towards fine granularity emotion |
| CN110083684B (en) * | 2019-04-24 | 2021-11-19 | 吉林大学 | Interpretable recommendation model for fine-grained emotion |
| CN110765769A (en) * | 2019-08-27 | 2020-02-07 | 电子科技大学 | Entity attribute dependency emotion analysis method based on clause characteristics |
| CN110765769B (en) * | 2019-08-27 | 2023-05-02 | 电子科技大学 | Clause feature-based entity attribute dependency emotion analysis method |
| CN112784011A (en) * | 2021-01-04 | 2021-05-11 | 南威软件股份有限公司 | Emotional problem processing method, device and medium based on CNN and LSTM |
| CN112784011B (en) * | 2021-01-04 | 2023-06-30 | 南威软件股份有限公司 | Emotion problem processing method, device and medium based on CNN and LSTM |
| CN112905736A (en) * | 2021-01-27 | 2021-06-04 | 郑州轻工业大学 | Unsupervised text emotion analysis method based on quantum theory |
| CN112905736B (en) * | 2021-01-27 | 2023-09-19 | 郑州轻工业大学 | Quantum theory-based unsupervised text emotion analysis method |
| CN115794988A (en) * | 2022-09-13 | 2023-03-14 | 广东美云智数科技有限公司 | Method, apparatus, and computer storage medium for extracting viewpoint of text |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103995803B (en) | 2017-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103995803B (en) | A kind of fine granularity text sentiment analysis method | |
| Mohammad et al. | Using hashtags to capture fine emotion categories from tweets | |
| Murphy | Linguistic representation and processing of copredication | |
| Kunneman et al. | Signaling sarcasm: From hyperbole to hashtag | |
| Barbieri et al. | Modelling sarcasm in twitter, a novel approach | |
| González-Ibánez et al. | Identifying sarcasm in twitter: a closer look | |
| US10642975B2 (en) | System and methods for automatically detecting deceptive content | |
| Han et al. | Knowledge of words: An interpretable approach for personality recognition from social media | |
| Flekova et al. | Personality profiling of fictional characters using sense-level links between lexical resources | |
| Ghag et al. | Comparative analysis of the techniques for sentiment analysis | |
| WO2016051551A1 (en) | Text generation system | |
| WO2014039897A1 (en) | System and method for mapping semiotic relationships | |
| CN104331394A (en) | Text classification method based on viewpoint | |
| Bellot et al. | INEX Tweet Contextualization task: Evaluation, results and lesson learned | |
| CN105183717A (en) | OSN user emotion analysis method based on random forest and user relationship | |
| Long et al. | Turbulent flow: A computational model of world literature | |
| Bloom | Sentiment analysis based on appraisal theory and functional local grammars | |
| Galitsky | Discovering rhetoric agreement between a request and response | |
| Helasvuo et al. | Choosing between zero and pronominal subject: modeling subject expression in the 1st person singular in Finnish conversation | |
| Snider | An exemplar model of syntactic priming | |
| Neviarouskaya et al. | Attitude sensing in text based on a compositional linguistic approach | |
| Malandrakis et al. | Sail: Sentiment analysis using semantic similarity and contrast features | |
| US20130226844A1 (en) | Content Summarizing and Search Method and System Using Thinking System | |
| Simaki et al. | Evaluating stance-annotated sentences from the Brexit Blog Corpus: A quantitative linguistic analysis | |
| Murphy et al. | Signals of contrastiveness: But, oppositeness, and formal similarity in parallel contexts |
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 |
Granted publication date: 20170704 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |