CN105095183A - Text emotional tendency determination method and system - Google Patents
Text emotional tendency determination method and system Download PDFInfo
- Publication number
- CN105095183A CN105095183A CN201410219390.1A CN201410219390A CN105095183A CN 105095183 A CN105095183 A CN 105095183A CN 201410219390 A CN201410219390 A CN 201410219390A CN 105095183 A CN105095183 A CN 105095183A
- Authority
- CN
- China
- Prior art keywords
- user
- emotion
- vocabulary
- text
- language material
- 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
Abstract
The present invention provides a text emotional tendency determination method and system. The text emotional tendency determination method comprises: a corpus acquisition step: acquiring user history text information within a certain time window as a corpus; a user personality characteristic determination step: determining personality characteristics of a user according to the corpus; an emotional vocabulary weight adjustment step: acquiring emotional vocabularies in a user text that needs to be determined and initial weights by using an emotional vocabulary dictionary to form an emotional vocabulary list, and performing adjustment on vocabulary weights in the emotional vocabulary list according to the personality characteristics of the user, determined in the user personality characteristic determination step; and a text emotional tendency determination step: according to the polarity of each vocabulary in the emotional vocabulary list and the adjusted weight of each vocabulary, determining emotional tendency of the user text that needs to be determined.
Description
Technical field
The user version Sentiment orientation that the present invention relates in technical field of data processing judges.More specifically, the present invention relates to a kind of text emotion based on user's character trait tendency determination methods and system.
Background technology
Along with SNS (SocialNetworkingServices, social network services) is constantly popular, people, by the platform such as microblogging, forum, issue various viewpoint to personage, event, product.In order to effectively process these information, finding the attitude suggestion of people, just needing to carry out text emotion analysis.Text emotion is inclined to, and have expressed the hobby of people to certain event, certain product.
But in the today in personalized epoch, the mode that everyone shows emotion and degree are different.Such as " fortunately " this word, some are represented that " good " then represents " generally " other people.Therefore how Judgment by emotion just objective reality can be done according to everyone different character trait, expression characteristic.
In order to solve the above-mentioned problem.First we look at art methods, and the deficiency of the method.
Patent Document 1 discloses a kind of sentiment analysis system and method, this system comprises corpus and sets up module, for setting up the training set needed for the identification of viewpoint sentence and Sentiment orientation analysis; Data prediction module, for carrying out pre-service to the sentence in training set; Viewpoint sentence identification module, adopts support vector machine classifier and Bayes classifier to carry out the identification of viewpoint sentence to pretreated sentence respectively, and carries out integrated process to the result of two sorters, obtain final classification results; And Sentiment orientation analysis module, based on support vector machine classifier and Bayes classifier direct front, negative and without viewpoint three class that pretreated sentence is divided into respectively, and by an integrated formula by integrated for the classification results of this support vector machine classifier and Bayes classifier, obtain the classification results of current sentence
Patent Document 2 discloses a kind of sentiment analysis method towards microblogging short text, the method comprises: step 1, and collection comprises the microblog data of designated key words stored in database; Step 2, carries out pre-service to microblog data; Step 3, loads relevant dictionary; Step 4, carries out subordinate sentence, filters out and does not comprise the sentence that user configures key word; Step 5, carries out participle, part-of-speech tagging to the sentence comprising key word; Step 6, utilizes syntactic analysis instrument to carry out interdependent syntactic analysis to the sentence comprising theme; Step 7, judges the polarity of each sentence comprising descriptor; Step 8, judged all comprise the polarity of the sentence of descriptor after, judge the emotion tendency of whole piece microblogging.
But all there is a common issue in the prior art: the character trait and the expression way that have ignored people itself, sentiment analysis carries out standardized calculation according to unified indiscriminate mode, the inevitable distortion of the Sentiment orientation drawn.
Prior art document
Patent documentation
Patent documentation 1:CN103034626A
Patent documentation 2:CN102663046A
Summary of the invention
The present invention researches and develops in view of the above problems, and object is to provide a kind of the text emotion tendency determination methods and the system that consider user's character trait, improves the accuracy that text emotion tendency judges.
One aspect of the present invention relates to a kind of text emotion tendency determination methods, and it is characterized in that comprising: language material obtains step, the user's history text information in certain hour window of obtaining is as language material; User's character trait determining step, judges the character trait of user according to described language material; Emotion vocabulary weight adjusting step, utilizing emotion vocabulary dictionary to obtain needs the emotion vocabulary in the user version judged and initial weight to form emotion word lists, and adjusts the term weight in described emotion word lists according to the character trait of the user judged in described user's character trait determining step; And text emotion tendency determining step, according to the polarity of each vocabulary in described emotion word lists and by adjusted weight, judge the described Sentiment orientation needing the user version judged.
In addition, preferred in the present invention: also to comprise text-processing step, in described text-processing step, the user version judged described language material or described needs carries out participle, part of speech judges, and the emotion vocabulary in utilize emotion vocabulary dictionary to obtain user version that described language material or described needs judge, in described user's character trait determining step, remit the character trait judging user according to the emotion word in described language material.
In addition, in the present invention preferably: described user's character trait determining step comprises: described language material is carried out the step sorted according to the time that text is issued; According to the time, cluster is carried out to described language material, and different clusters is kept at the step in different language material set respectively; Language material in language material set described in each is analyzed, judges the step of the user's character trait for current language material set; And COMPREHENSIVE CALCULATING is carried out to user's character trait of whole described language material set, thus obtain the step of the character trait of final user.
In addition, in the present invention preferably: when carrying out COMPREHENSIVE CALCULATING to user's character trait of whole described language material set, following mathematical expression is utilized to adjust the weights W of user's character trait,
Wherein, T is the constant representing damped cycle, and t is the time interval, and e is regulation constant.
It is preferred in the present invention: in described emotion vocabulary weight adjusting step, to utilize the weight S of following mathematical expression to emotion vocabulary to adjust,
Wherein, a be greater than 1 empirical parameter, S
0be the initial weight of emotion vocabulary, x is the user's character trait after quantizing, and y is the polarity of emotion vocabulary.
In addition, preferred in the present invention: in described text emotion tendency determining step, when the absolute value of the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is not more than given threshold value beta, the described Sentiment orientation of the user version judged that needs is judged as neutrality, when the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is greater than β, the described Sentiment orientation of the user version judged that needs is judged as positivity, when the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is less than-β, the described Sentiment orientation of the user version judged that needs is judged as negativity.
Another aspect of the present invention relates to a kind of text emotion tendency judgement system, it is characterized in that comprising: language material acquisition unit, and the user's history text information in certain hour window of obtaining is as language material; User's character trait judging unit, judges the character trait of user according to described language material; Emotion vocabulary weight adjustment unit, utilizing emotion vocabulary dictionary to obtain needs the emotion vocabulary in the user version judged and initial weight to form emotion word lists, and adjusts the term weight in described emotion word lists according to the character trait of the user judged by described user's character trait judging unit; And text emotion tendency judging unit, according to the polarity of each vocabulary in described emotion word lists and by adjusted weight, judge the described Sentiment orientation needing the user version judged.
In addition, preferred in the present invention: also to comprise text-processing unit, in described text-processing unit, the user version judged described language material or described needs carries out participle, part of speech judges, and the emotion vocabulary in utilize emotion vocabulary dictionary to obtain user version that described language material or described needs judge, described user's character trait judging unit remits the character trait judging user according to the emotion word in described language material.
Invention effect
According to the present invention, can judge that the personality of user, mood, expression style (i.e. user's character trait) carry out sentiment analysis again according to the history language material of user.The work such as information recommendation (as commodity, good friend, news etc.), evaluation accurately can be carried out according to the method.
Accompanying drawing explanation
Accompanying drawing is used for doing further understanding to the present invention, forms a part for instructions, is used from and carries out detailed explanation to the present invention, be not construed as limiting the invention with preferred embodiment one.Wherein:
Fig. 1 is the schematic block diagram of the text emotion tendency judgement system that the present invention relates to.
Fig. 2 is the main flow chart of the text emotion tendency determination methods that the present invention relates to.
Fig. 3 is that the present invention is for judging the process flow diagram of user's character trait.
Fig. 4 is that the present invention is for adjusting the process flow diagram of emotion term weight.
Embodiment
Come below with reference to the accompanying drawings to be described in detail to embodiments of the present invention, but the present invention is not limited to this embodiment.In addition, in following description of the present invention, the specific descriptions to known function and configuration will be omitted, to avoid making theme of the present invention unclear.
As shown in Figure 1, the text emotion tendency judgement system of present embodiment comprises: language material acquisition unit 101, user's character trait judging unit 102, emotion vocabulary weight adjustment unit 103, text emotion tendency judging unit 104, user interface section 105 and text-processing unit 106.
Language material acquisition unit 101 is collected in the history text information of user in certain hour window as language material.Time window can be sky, Yue Deng unit, also can be the chronomere that user is arranged voluntarily.The API provided by website carries out content crawl, also can carry out text collection by web crawlers.The network text information acquired is saved to this locality and is stored as corpus.Sorted the time that language material is issued according to text, then carry out cluster according to the time to text, the rule of cluster is that to be gathered by the text that similar time section is issued be a class, and the text of different time cluster is left in different set.Afterwards, text bunch in each time set (language material set) is analyzed, existing Text Mining Technology can be utilized and in conjunction with psychology relevant knowledge, mode as by supervised learning sets up a binary classification model, utilize this model to judge the tendentiousness of user's personality, actively whether optimism is still remained passive and pessimistic etc.Certainly according to psychologic theory, the personality of people is not limited to two classes, and the K sorting technique of machine learning can be utilized equally to carry out modeling.Here we have done simplification to problem.
Text-processing unit 106 is for carrying out word segmentation processing to the sentence in corpus and finding neologisms by prior art.Utilize related software, as ICTCLAS etc. carries out participle, part of speech judgement to sentence; Meanwhile, the sentiment dictionary both deposited is utilized to filter the emotion vocabulary occurred in user version.
Emotion vocabulary weight adjustment unit 103, utilizing emotion vocabulary dictionary to obtain needs the emotion word in the user version detected and initial weight to form emotion word lists, then readjusts the term weight in this list according to the character trait of user.
here, a is the individual empirical parameter being greater than 1, S
0it is the initial weight (forward is greater than zero, and negative sense is less than zero) of emotion vocabulary Sentiment orientation; X is the user's character trait after quantizing, and 0 represents common, and 1 represents positive forward, and-1 represents passive negative sense; Y is the polarity (non-weight) of the original emotion vocabulary after quantizing, and forward vocabulary is 1, and negative sense vocabulary is-1.By this formula, we can obtain, the people of consistent actively forward if negative sense vocabulary then this vocabulary will increase weight, if but this people has said the vocabulary of forward, instead this vocabulary can be reduced weight.Vice versa.
Text emotion tendency judging unit 104, judges the polarity of each word in emotion word lists, then obtains the Sentiment orientation (polarity) of the user version needing to judge according to COMPREHENSIVE CALCULATING formula.COMPREHENSIVE CALCULATING formula is
Wherein, S is net result, and Pos{W} is the weight sum of all forward vocabulary, and Neg{W} is the weight sum of all negative sense vocabulary, and Neu{W} is the weight sum of all neutral vocabulary.Here, in order to ensure the accuracy judged, introduce threshold values β during calculating, this value is an empirical value, when the absolute value of the difference of positive and negative contribution rate is not more than β, result of calculation is set to neutrality.Especially, when threshold values β is 0, then direct by positive and negative that class vocabulary larger to weight decision polarity.
User interface section 105, the configuration that main completing user carries out system, input and show the interface of neologisms and polarity thereof.
Referring to Fig. 2 and Fig. 3, the flow process of the text emotion tendency determination methods of present embodiment is described.
(1) first, the character trait (step S201-S202) judging user is needed.
Concrete steps are as follows:
(1.1) language material obtains step, collects the history text information (step S201) of user.Be collected in a timing and ask that the history text information of user in window is as language material.Time window can be sky, Yue Deng unit, also can be the chronomere that user is arranged voluntarily.The API provided by website carries out content crawl, also can carry out text collection by web crawlers.The network text information acquired is saved to this locality and is stored as corpus.
(1.2) user's character trait determining step, analyzes history text, judges user's character trait (step S202).
Following step can be subdivided into again for this step:
(1.2.1) according to the time issued, (step S301) is sorted to history text information.Although consider that each user has metastable character trait but have different anxious state of mind in regular hour section, these fluctuations also can affect in the result of sentiment analysis.
(1.2.2) according to the time, cluster is carried out to text and different cluster texts is kept at (step S302) in different set respectively.The object of cluster is exactly text in the time period far away for interval in text set in similar time section will be separated.Such as user all posts every day before 2 months and the amount of posting is intensive, and afterwards 1 month relatively quiet, and start now active.Therefore the history text of user can be clustered into 2 class amounts, before the February, one present.The concrete grammar of cluster has multiple, such as K-Mean etc., also can define a density thresholds, is classified as a class when density of posting is greater than threshold values.Especially, when user evenly posts all the time, do not have the obvious time interval, in this case, only produce a cluster, all texts are all in the middle of this cluster.Certain the present invention is not limited to these object lessons.
(1.2.3) carry out user characteristics modeling by the method for text mining, and according to this model, (step S303) is classified to user's character trait.Respectively mining analysis is carried out to the text in different time set, the result weight of forward personality is put and is greater than 0 and negative sense personality result is less than 0 neutral personality is 0.That is, text bunch in the set of each time of previous step is analyzed, existing Text Mining Technology can be utilized and in conjunction with psychology relevant knowledge, mode as by supervised learning sets up a binary classification model, utilize this model to judge the tendentiousness of user's personality, actively whether optimism is still remained passive and pessimistic etc.Certainly according to psychologic theory, the personality of people is not limited to two classes, and the K sorting technique of machine learning can be utilized equally to carry out modeling.Such as, suppose that the personality of forward represents with 1, representing with 1 of negative sense.
(1.2.4) after, we do Macro or mass analysis (step S304) to the character trait of different time cluster.Consider undulatory property and the continuity of mood.Time nearer mood more can represent true emotional instantly.Therefore, the principle gathered be from present time more close to the more large cluster weight far away of cluster weight less, that is:
wherein T is constant, represents damped cycle, and t is the time interval, that is, the interval of the generation time distance current time of user version, and Ke Yi, the moon etc. are unit, and e is math constant, that is, the truth of a matter of natural logarithm.Then by the mode of algebraic sum,
determine final user's personality polarity: if be greater than 0, be forward, be less than 0 for negative sense, equal zero then for neutral.
(2) after acquisition user character trait, the weight of emotion vocabulary is adjusted (step S203).
Concrete step is as follows:
(2.1) from emotion vocabulary dictionary, obtain the original emotion weight (step 401) of vocabulary.Carry out text-processing to the text of current Water demand, the emotion vocabulary found out wherein generates emotion word lists.This process can utilize related software, as ICTCLAS etc. carries out participle, part of speech judgement to sentence, and utilizes the emotion vocabulary dictionary both deposited to filter the emotion vocabulary occurred in user version, thus obtains emotion word lists.About the definition of emotion vocabulary dictionary, can with reference to table 1 below.For there being adverbial word to describe before emotion word, the weight of this word can be adjusted according to the method for product according to the coefficient of adverbial word.Adverbial word vocabulary can with reference to table 2 below.For there being negative word before emotion word, its weight is multiplied by-1.
[table 1]
Sentiment dictionary
Word | Polarity | Weight |
Good | Just | 1 |
Amiable | Just | 1 |
Succinctly | Just | 1 |
... | ... | |
Sorrow | Negative | -1 |
Bad | Negative | -1 |
Profound | Negative | -1 |
Jiong | Negative | -1 |
... | ... |
[table 2]
Adverbial word subscale
Magnitude | Degree adverb | Coefficient |
Maximum dose | , the most too, very .. | 3 |
A large amount | More, more, very, especially ... | 2.5 |
Middle amount | Comparatively, not too, not bery ... | 2 |
Low amounts | A little, slightly ... | 1.5 |
(2.2) according to the character trait of user, emotion term weight is adjusted (step S402, S403, S404).The principle of adjustment be people for those positive forwards if forward vocabulary then this word weight reduce, say negative sense vocabulary then this word weight increase.Vice versa.Therefore have
here, a is the individual empirical parameter being greater than 1, S
0it is the initial weight (forward is greater than zero, and negative sense is less than zero) of emotion word Sentiment orientation; X is the user's personality after quantizing, and 0 represents common, and 1 represents positive forward, and one 1 represent passive negative sense; Y is the polarity (non-weight) of the original emotion vocabulary after quantizing, and forward vocabulary is 1, and negative sense vocabulary is 1.
(3) after have updated emotion term weight, the final Sentiment orientation (step S204) that will detect text is calculated.COMPREHENSIVE CALCULATING formula is
Wherein, S is net result, and Pos{W} is the weight sum of all forward vocabulary, and Neg{W} is the weight sum of all negative sense vocabulary, and Neu{W} is the weight sum of all neutral vocabulary.Here, in order to ensure the accuracy judged, introduce threshold values 13 during calculating, this value is an empirical value, when the absolute value of the difference of positive and negative contribution rate is not more than β, result of calculation is set to neutrality.Especially, when threshold values β is 0, then direct by positive and negative that class vocabulary larger to weight decision polarity.
(4) last, the result of judgement is returned.
Below in conjunction with specific embodiment, embodiment is further elaborated.It should be noted that following embodiment just the present invention and the object lesson enumerated for convenience of explanation, the present invention is not limited to these embodiments.
(embodiment 1)
Embodiment 1 is the object lesson that user initiatively have input the emotion word that will detect.
(routine 1-1) microblog users first.To certain book review: " although this this book publishing fineness, but dull in content boring, it's a sheer waste of time! " (issue in November)
1. user characteristics judges.Collect his content of microblog over the past half year recently.Example:
1. " the Wang Xing people of neighbour's is simply too lovely [blocking her] " (attached dog photo); September is issued
2. " Beijing in September is very comfortable, makes us happy "; September is issued
3. " every day is all happy, transmits positive energy "; September is issued
........
4. " report is submitted to Thursday, and a word is not also write [going mad] "; November is issued
5. " scold by boss, very not well "; November is issued
......
Carry out cluster according to the time, we obtain two set (representative of set element microblogging id number):
SetA={1,2,3},SetB={4,5}
Respectively these two cluster texts are analyzed.Model is set up has measure of supervision to carry out with above-mentioned.This process is not repeating, and directly classifies with this model.Obviously, SetA is positive forward, represents with 1; SetB is passive negative sense, represents with-1.Again according to time factor adjustment,
in order to convenience of calculation, if T is 1, t is time interval in units of the moon.Therefore, concerning its feature weight SetA be:
and concerning SetB:
So final W=Wa+Wb < 0 of the character trait of user, namely user is negative sense character trait.
2. emotion vocabulary weight adjusting.We carry out participle and part of speech extraction with participle instrument to the current text that will detect of user, and filter with emotion word dictionary.The emotion word that can obtain in text is: C={ is exquisite, uninteresting, boring, waste }.Obtain emotion word lists:
[table 3]
Emotion word word lists
Word | Polarity | Weight |
Exquisite | Just (1) | 1 |
Uninteresting | Negative (-1) | -1 |
Boring | Negative (-1) | -1 |
Waste | Negative (-1) | -1 |
Checking in text does not have other adverbial words, negative word to modify.
Utilize character trait to adjust the value in above list, can obtain:
3. the Judgment by emotion of text.For in previous step, assuming that we get empirical value a=1.5, so final result of calculation S=a-a
∧(-1)-a
∧(-1)-a
∧(-1)=1.5-1.5
∧(-1)-1.5
∧(-1)-1.5
∧(-1)=-0.5 < 0, therefore the final Judgment by emotion of text is negativity.
Text emotion tendency determination methods of the present invention comprises: user characteristics detecting step, by the user's history text information in selection certain hour window as language material, classifies, obtain user's character trait according to the expression characteristic in this language material; Emotion term weight assignment procedure again, utilizing emotion word dictionary to obtain needs the emotion word in the user version detected and initial weight to form emotion word lists, then readjusts the term weight in these lists according to the character trait of user; Text emotion tendency detecting step, is obtained the polarity of each vocabulary in word lists, gathers the feeling polarities obtaining text by previous step.
Text emotion of the present invention tendency judgement system comprises: user characteristics judge module, and it is collected in user's history text information in certain hour window as language material, classifies, obtain user's character trait according to the expression characteristic in this language material; Text processing module, it cuts the functions such as word, part of speech judgement, emotion word filtration to the sentence in corpus; Emotion word weight adjusting module, utilizing emotion word dictionary to obtain needs the emotion word in the user version detected and initial weight to form emotion word lists, then readjusts the term weight in these lists according to the character trait of user; Text emotion analyzes judge module, and according to polarity and the weight of word each in word lists, COMPREHENSIVE CALCULATING formula obtains the feeling polarities of text.
The present invention can be applicable to the aspect such as opining mining, product evaluation, and then can carry out the recommendation of relevant information (as commodity, good friend, news etc.).
According to the present invention, can judge that the personality of user, mood, expression style (i.e. user's character trait) carry out sentiment analysis again according to the history language material of user.Therefore, it is possible to carry out the work such as information recommendation (as commodity, good friend, news etc.), evaluation accurately.
Above embodiments of the present invention and specific embodiment are illustrated, but the present invention is not defined in above-mentioned embodiment.In addition, for above-mentioned embodiment, do not departing from the scope of the meaning shown in the word described in purport of the present invention and claims, implementing the various distortion that can expect of those skilled in the art and obtain variation and be also contained in the present invention.
Claims (8)
1. a text emotion tendency determination methods, is characterized in that comprising:
Language material obtains step, and the user's history text information in certain hour window of obtaining is as language material;
User's character trait determining step, judges the character trait of user according to described language material;
Emotion vocabulary weight adjusting step, utilizing emotion vocabulary dictionary to obtain needs the emotion vocabulary in the user version judged and initial weight to form emotion word lists, and adjusts the term weight in described emotion word lists according to the character trait of the user judged in described user's character trait determining step; And
Text emotion tendency determining step, according to the polarity of each vocabulary in described emotion word lists and by adjusted weight, judges the described Sentiment orientation needing the user version judged.
2. text emotion tendency determination methods according to claim 1, is characterized in that,
Also comprise text-processing step, in described text-processing step, the user version judged described language material or described needs carries out participle, part of speech judges, and the emotion vocabulary in utilize emotion vocabulary dictionary to obtain user version that described language material or described needs judge
In described user's character trait determining step, remit the character trait judging user according to the emotion word in described language material.
3. text emotion tendency determination methods according to claim 1, is characterized in that,
Described user's character trait determining step comprises:
Described language material is carried out the step sorted according to the time that text is issued;
According to the time, cluster is carried out to described language material, and different clusters is kept at the step in different language material set respectively;
Language material in language material set described in each is analyzed, judges the step of the user character trait relevant to current language material set; And
COMPREHENSIVE CALCULATING is carried out to user's character trait of whole described language material set, thus obtains the step of the character trait of final user.
4. text emotion tendency determination methods according to claim 3, is characterized in that,
When carrying out COMPREHENSIVE CALCULATING to user's character trait of whole described language material set, following mathematical expression is utilized to adjust the weights W of user's character trait,
Wherein, T is the constant representing damped cycle, and t is the time interval, and e is regulation constant.
5. text emotion tendency determination methods according to claim 1, is characterized in that,
In described emotion vocabulary weight adjusting step, the weight S of following mathematical expression to emotion vocabulary is utilized to adjust,
Wherein, a be greater than 1 empirical parameter, S
0be the initial weight of emotion vocabulary, x is the user's character trait after quantizing, and y is the polarity of emotion vocabulary.
6. text emotion tendency determination methods according to claim 1, is characterized in that,
In described text emotion tendency determining step,
When the absolute value of the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is not more than given threshold value beta, the described Sentiment orientation of the user version judged that needs is judged as neutrality,
When the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is greater than β, the described Sentiment orientation of the user version judged that needs is judged as positivity,
When the difference of the contribution rate of all forward vocabulary in described emotion word lists and the contribution rate of all negative sense vocabulary is less than-β, the described Sentiment orientation of the user version judged that needs is judged as negativity.
7. a text emotion tendency judgement system, is characterized in that comprising:
Language material acquisition unit, the user's history text information in certain hour window of obtaining is as language material;
User's character trait judging unit, judges the character trait of user according to described language material;
Emotion vocabulary weight adjustment unit, utilizing emotion vocabulary dictionary to obtain needs the emotion vocabulary in the user version judged and initial weight to form emotion word lists, and adjusts the term weight in described emotion word lists according to the character trait of the user judged by described user's character trait judging unit; And
Text emotion tendency judging unit, according to the polarity of each vocabulary in described emotion word lists and by adjusted weight, judges the described Sentiment orientation needing the user version judged.
8. text emotion tendency judgement system according to claim 7, is characterized in that,
Also comprise text-processing unit, in described text-processing unit, the user version judged described language material or described needs carries out participle, part of speech judges, and the emotion vocabulary in utilize emotion vocabulary dictionary to obtain user version that described language material or described needs judge
Described user's character trait judging unit remits the character trait judging user according to the emotion word in described language material.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410219390.1A CN105095183A (en) | 2014-05-22 | 2014-05-22 | Text emotional tendency determination method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410219390.1A CN105095183A (en) | 2014-05-22 | 2014-05-22 | Text emotional tendency determination method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105095183A true CN105095183A (en) | 2015-11-25 |
Family
ID=54575655
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410219390.1A Pending CN105095183A (en) | 2014-05-22 | 2014-05-22 | Text emotional tendency determination method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105095183A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105912563A (en) * | 2016-03-23 | 2016-08-31 | 北京数字跃动科技有限公司 | Method of giving machines artificial intelligence learning based on knowledge of psychology |
CN107133836A (en) * | 2017-03-22 | 2017-09-05 | 无锡中科富农物联科技有限公司 | A kind of adaptive weighting combined recommendation algorithm |
CN107480136A (en) * | 2017-08-02 | 2017-12-15 | 陈雷 | It is a kind of to be applied to the method that feeling curve is analyzed in screen play |
CN107526831A (en) * | 2017-09-04 | 2017-12-29 | 华为技术有限公司 | A kind of natural language processing method and apparatus |
CN107645559A (en) * | 2017-09-30 | 2018-01-30 | 广东美的制冷设备有限公司 | Household electrical appliances information-pushing method, server, mobile terminal and storage medium |
CN108090097A (en) * | 2016-11-23 | 2018-05-29 | 北京国双科技有限公司 | Text handling method and device |
CN108319588A (en) * | 2018-02-13 | 2018-07-24 | 北京易真学思教育科技有限公司 | Text emotion analysis system and method, storage medium |
CN109254993A (en) * | 2017-07-07 | 2019-01-22 | 北京掌沃云视媒文化传媒有限公司 | A kind of text based personality data analysing method and system |
CN109446378A (en) * | 2018-11-08 | 2019-03-08 | 北京奇艺世纪科技有限公司 | Information recommendation method, Sentiment orientation determine method and device and electronic equipment |
US10311069B2 (en) | 2016-06-02 | 2019-06-04 | International Business Machines Corporation | Sentiment normalization using personality characteristics |
CN110263344A (en) * | 2019-06-25 | 2019-09-20 | 名创优品(横琴)企业管理有限公司 | A kind of text emotion analysis method, device and equipment based on mixed model |
CN111105776A (en) * | 2018-10-26 | 2020-05-05 | 财团法人资讯工业策进会 | Audio playing device and playing method thereof |
WO2021114841A1 (en) * | 2020-05-14 | 2021-06-17 | 平安科技(深圳)有限公司 | User report generating method and terminal device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020163500A1 (en) * | 2001-04-23 | 2002-11-07 | Griffith Steven B. | Communication analyzing system |
CN101639824A (en) * | 2009-08-27 | 2010-02-03 | 北京理工大学 | Text filtering method based on emotional orientation analysis against malicious information |
CN101894102A (en) * | 2010-07-16 | 2010-11-24 | 浙江工商大学 | Method and device for analyzing emotion tendentiousness of subjective text |
CN103150367A (en) * | 2013-03-07 | 2013-06-12 | 宁波成电泰克电子信息技术发展有限公司 | Method for analyzing emotional tendency of Chinese microblogs |
JP2013257677A (en) * | 2012-06-12 | 2013-12-26 | Nec Corp | Event detection device, event detection method and event detection program |
CN103714071A (en) * | 2012-09-29 | 2014-04-09 | 株式会社日立制作所 | Label emotional tendency quantifying method and label emotional tendency quantifying system |
-
2014
- 2014-05-22 CN CN201410219390.1A patent/CN105095183A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20020163500A1 (en) * | 2001-04-23 | 2002-11-07 | Griffith Steven B. | Communication analyzing system |
CN101639824A (en) * | 2009-08-27 | 2010-02-03 | 北京理工大学 | Text filtering method based on emotional orientation analysis against malicious information |
CN101894102A (en) * | 2010-07-16 | 2010-11-24 | 浙江工商大学 | Method and device for analyzing emotion tendentiousness of subjective text |
JP2013257677A (en) * | 2012-06-12 | 2013-12-26 | Nec Corp | Event detection device, event detection method and event detection program |
CN103714071A (en) * | 2012-09-29 | 2014-04-09 | 株式会社日立制作所 | Label emotional tendency quantifying method and label emotional tendency quantifying system |
CN103150367A (en) * | 2013-03-07 | 2013-06-12 | 宁波成电泰克电子信息技术发展有限公司 | Method for analyzing emotional tendency of Chinese microblogs |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105912563A (en) * | 2016-03-23 | 2016-08-31 | 北京数字跃动科技有限公司 | Method of giving machines artificial intelligence learning based on knowledge of psychology |
CN105912563B (en) * | 2016-03-23 | 2019-04-02 | 北京数字跃动科技有限公司 | A method of the artificial intelligence learning of machine is assigned based on psychological knowledge |
US11106687B2 (en) | 2016-06-02 | 2021-08-31 | International Business Machines Corporation | Sentiment normalization using personality characteristics |
US10311069B2 (en) | 2016-06-02 | 2019-06-04 | International Business Machines Corporation | Sentiment normalization using personality characteristics |
CN108090097A (en) * | 2016-11-23 | 2018-05-29 | 北京国双科技有限公司 | Text handling method and device |
CN107133836A (en) * | 2017-03-22 | 2017-09-05 | 无锡中科富农物联科技有限公司 | A kind of adaptive weighting combined recommendation algorithm |
CN109254993A (en) * | 2017-07-07 | 2019-01-22 | 北京掌沃云视媒文化传媒有限公司 | A kind of text based personality data analysing method and system |
CN109254993B (en) * | 2017-07-07 | 2021-06-01 | 掌沃云科技(北京)有限公司 | Text-based character data analysis method and system |
CN107480136A (en) * | 2017-08-02 | 2017-12-15 | 陈雷 | It is a kind of to be applied to the method that feeling curve is analyzed in screen play |
CN107480136B (en) * | 2017-08-02 | 2020-07-03 | 逄泽沐风 | Method applied to emotional curve analysis in movie script |
CN107526831A (en) * | 2017-09-04 | 2017-12-29 | 华为技术有限公司 | A kind of natural language processing method and apparatus |
US11630957B2 (en) | 2017-09-04 | 2023-04-18 | Huawei Technologies Co., Ltd. | Natural language processing method and apparatus |
CN107645559A (en) * | 2017-09-30 | 2018-01-30 | 广东美的制冷设备有限公司 | Household electrical appliances information-pushing method, server, mobile terminal and storage medium |
CN107645559B (en) * | 2017-09-30 | 2020-10-09 | 广东美的制冷设备有限公司 | Household appliance information pushing method, server, mobile terminal and storage medium |
CN108319588A (en) * | 2018-02-13 | 2018-07-24 | 北京易真学思教育科技有限公司 | Text emotion analysis system and method, storage medium |
CN111105776A (en) * | 2018-10-26 | 2020-05-05 | 财团法人资讯工业策进会 | Audio playing device and playing method thereof |
CN109446378A (en) * | 2018-11-08 | 2019-03-08 | 北京奇艺世纪科技有限公司 | Information recommendation method, Sentiment orientation determine method and device and electronic equipment |
CN110263344B (en) * | 2019-06-25 | 2022-04-19 | 创优数字科技(广东)有限公司 | Text emotion analysis method, device and equipment based on hybrid model |
CN110263344A (en) * | 2019-06-25 | 2019-09-20 | 名创优品(横琴)企业管理有限公司 | A kind of text emotion analysis method, device and equipment based on mixed model |
WO2021114841A1 (en) * | 2020-05-14 | 2021-06-17 | 平安科技(深圳)有限公司 | User report generating method and terminal device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105095183A (en) | Text emotional tendency determination method and system | |
Al Asad et al. | Depression detection by analyzing social media posts of user | |
CN105468713A (en) | Multi-model fused short text classification method | |
Desai et al. | Techniques for sentiment analysis of Twitter data: A comprehensive survey | |
CN103699626B (en) | Method and system for analysing individual emotion tendency of microblog user | |
CN103399891B (en) | Method for automatic recommendation of network content, device and system | |
Han et al. | Knowledge of words: An interpretable approach for personality recognition from social media | |
CN109829166B (en) | People and host customer opinion mining method based on character-level convolutional neural network | |
CN102096680A (en) | Method and device for analyzing information validity | |
CN105138577B (en) | Big data based event evolution analysis method | |
CN103207860A (en) | Method and device for extracting entity relationships of public sentiment events | |
CN112199608A (en) | Social media rumor detection method based on network information propagation graph modeling | |
CN108733675B (en) | Emotion evaluation method and device based on large amount of sample data | |
CN103473262A (en) | Automatic classification system and automatic classification method for Web comment viewpoint on the basis of association rule | |
JP2006293767A (en) | Sentence categorizing device, sentence categorizing method, and categorization dictionary creating device | |
Ramírez-Tinoco et al. | A brief review on the use of sentiment analysis approaches in social networks | |
CN110457711B (en) | Subject word-based social media event subject identification method | |
Ahmed et al. | A novel approach for Sentimental Analysis and Opinion Mining based on SentiWordNet using web data | |
Dikiyanti et al. | Sentiment analysis and topic modeling of BPJS Kesehatan based on twitter crawling data using Indonesian Sentiment Lexicon and Latent Dirichlet Allocation algorithm | |
CN111309864A (en) | User group emotional tendency migration dynamic analysis method for microblog hot topics | |
Samonte et al. | Sentence-level sarcasm detection in English and Filipino tweets | |
WO2019235446A1 (en) | Degree of difficulty estimating device, and degree of difficulty estimating model learning device, method, and program | |
CN111859955A (en) | Public opinion data analysis model based on deep learning | |
Fritsche et al. | Deciphering professional forecasters' stories: Analyzing a corpus of textual predictions for the German economy | |
CN103793478A (en) | Online theme modeling method on basis of theme heredity |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 |
Application publication date: 20151125 |
|
WD01 | Invention patent application deemed withdrawn after publication |