CN105447036B - A kind of social media information credibility evaluation method and device based on opining mining - Google Patents
A kind of social media information credibility evaluation method and device based on opining mining Download PDFInfo
- Publication number
- CN105447036B CN105447036B CN201410436605.5A CN201410436605A CN105447036B CN 105447036 B CN105447036 B CN 105447036B CN 201410436605 A CN201410436605 A CN 201410436605A CN 105447036 B CN105447036 B CN 105447036B
- Authority
- CN
- China
- Prior art keywords
- information
- assessed
- social media
- media information
- viewpoint
- 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.)
- Active
Links
Abstract
The embodiment of the invention discloses a kind of social media information credibility evaluation method and device based on opining mining, method include: to obtain information to be assessed；Calculate the uncertain score of every information to be assessed；Calculate the confidence level of the publisher of every information to be assessed；It counts and supports ratio shared by opinion in the comment of every information to be assessed；By the uncertain score of every information to be assessed, support ratio shared by opinion inputs in trained quantitative appraisement model in advance to calculate in the comment of the confidence level of the publisher of every information to be assessed and every information to be assessed, the output of the quantitative appraisement model is the reliability order of every information to be assessed.The embodiment of the present invention can carry out accurate evaluation to the confidence level of social media information.
Description
Technical field
The present invention relates to fields of communication technology, and in particular to a kind of social media information confidence level based on opining mining is commented
Estimate method and device.
Background technique
With the development of second-generation internet WEB2.0 technology and universal, various types of social medias are (such as: microblogging, micro-
Letter, Twitter etc.) continue to bring out and profoundly change people's publication, obtain, exchange, expressing information or viewpoint mode.It is special
It is not with the maturation of ng mobile communication and being widely used for Intelligent mobile equipment, social media has become people
The indispensable platform for being used to sharing information and expressing viewpoint in daily life.But due to the content on the platform mainly by
The spontaneous creation of large quantities of netizens is extracted, so false, unreliable information is widely present.How social media letter is automatically assessed
The confidence level of breath will generate direct influence to the effect of the application systems such as subsequent information recommendation, market survey, automatic question answering.
The information credibility that the prior art provides is analyzed mainly for a certain specific area, certain types of data, such as
Biomedicine experiment report, newswire, wikipedia etc..By taking the reliability assessment of biomedicine experiment report as an example, due to
Such data has fixed structure and mode, it is possible to different features is easily extracted, especially for some class
Topic has more associated laboratory reports, it is possible to a low credibility to identify by the mutual verifying between laboratory report
Laboratory report.And the confidence level of wikipedia information is mainly to be characterized by the modification record of information.
I.e. early stage information credibility analysis tool primarily to the data of particular structured and design, do not examine
The text information considered in the data structure feature and language communicative habits, especially social media of social media information itself is
A kind of non-structured data, the processing of such data it is serious depend on natural language processing technique: such as semantic analysis and
Sentiment analysis etc., so the technology of this kind of early stage is not particularly suited for the reliability assessment of social media information.Therefore, it is necessary to mention
The confidence level of social media information is assessed for a kind of new method.
Summary of the invention
In view of this, the present invention provides a kind of social media information credibility evaluation method and dress based on opining mining
It sets, accurate evaluation can be carried out to the confidence level of social media information.
In a first aspect, the social media information credibility evaluation method provided in an embodiment of the present invention based on opining mining,
Include:
Obtain information to be assessed；
Calculate the uncertain score of every information to be assessed；
Calculate the confidence level of the publisher of every information to be assessed；
It counts and supports ratio shared by opinion in the comment of every information to be assessed；
By the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every to
It assesses and supports ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of information, it is described
The output of quantitative appraisement model is the reliability order of every information to be assessed.
With reference to first aspect, in the first embodiment of first aspect, before obtaining information to be assessed, the side
Method further include:
Construct theme dictionary relevant to current subject under discussion；
Each descriptor in the theme dictionary and each emotion word combination in emotion dictionary are formed into viewpoint word pair；
Obtain social media information relevant to current subject under discussion；
According to each viewpoint word pair and the similarity of every social media information and each viewpoint word pair and every social matchmaker
The viewpoint value of every social media information of similarity calculation of the comment of body information；
The social media information that viewpoint value is less than preset threshold is filtered, using remaining social media information as described to be evaluated
Estimate information.
In the first embodiment with reference to first aspect, in second of embodiment of first aspect, the building
Theme dictionary relevant to current subject under discussion specifically includes:
Social media information relevant to current subject under discussion is searched in social networks；
It extracts the keyword in the social media information and counts the frequency that each keyword occurs；
The theme dictionary is constructed as descriptor according to the keyword that the sequence of frequency from high to low chooses preset quantity.
In the first embodiment with reference to first aspect, in the third embodiment of first aspect, the basis
Each viewpoint word pair and the similarity of every social media information and the comment of each viewpoint word pair and every social media information
The viewpoint value of every social media information of similarity calculation specifically include:
The descriptor and the similarity of each keyword in a social media information for calculating a viewpoint word centering, mention
Take similarity maximum value a；Calculate each pass in the descriptor of the viewpoint word centering and the comment of the social media information
The similarity of keyword extracts similarity maximum value x；
The emotion word of the viewpoint word centering and the similarity of each emotion word in the social media information are calculated, is mentioned
Take similarity maximum value b；Calculate each feelings in the emotion word of the viewpoint word centering and the comment of the social media information
Feel the similarity of word, extracts similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 less than 1, institute
The similarity for stating viewpoint word pair and the comment of the social media information is s2=μ x+ (1- μ) y, and μ is greater than 0 less than 1；
By the viewpoint word pair and the similarity of the social media information and the viewpoint word pair and the social media
The similarity of the comment of information is added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained into all viewpoint subvalues of the social media information to same processing is done, by all sights
Point subvalue is cumulative to obtain the viewpoint value of the social media information, and so on, obtain the viewpoint value of each social media information.
With reference to first aspect or second of embodiment of the first embodiment of first aspect or first aspect, or
The third embodiment of first aspect, in the 4th kind of embodiment of first aspect, every information to be assessed of the calculating
Uncertain score include:
Determine the classification for the uncertain content for including in every information to be assessed；
Calculate the category score that the every class for including in every information to be assessed does not know content；
The category score that the every class for including in every information to be assessed does not know content is added up multiplied by after preset weight
Obtain the uncertain score of every information to be assessed.
With reference to first aspect or second of embodiment of the first embodiment of first aspect or first aspect, or
The third embodiment of first aspect, in the 5th kind of embodiment of first aspect, by every information to be assessed not
Certainty score is supported shared by opinion in the comment of the confidence level of the publisher of every information to be assessed and every information to be assessed
Ratio input and calculated in trained quantitative appraisement model in advance during, the uncertainty of the information to be assessed
Score is higher, and the confidence level of the information to be assessed is lower；The confidence level of the publisher of the information to be assessed is lower, it is described to
The confidence level for assessing information is lower；In the comment of the information to be assessed support opinion shared by ratio it is smaller, and/or with when
Between variation, support ratio shared by opinion smaller and smaller in the comment of the information to be assessed, the information to be assessed can
Reliability is lower.
Second aspect, the social media information reliability assessment device provided in an embodiment of the present invention based on opining mining,
Include:
First acquisition unit, for obtaining information to be assessed；
First computing unit, for calculating the uncertain score of every information to be assessed；
Second computing unit calculates the confidence level of the publisher of every information to be assessed；
Statistic unit supports ratio shared by opinion in the comment for counting every information to be assessed；
Reliability assessment unit, for by the uncertain score of every information to be assessed, the hair of every information to be assessed
Support that the input of ratio shared by opinion in advance comment by trained quantization in the comment of the confidence level of cloth person and every information to be assessed
Estimate in model and calculated, the output of the quantitative appraisement model is the reliability order of every information to be assessed.
In conjunction with second aspect, in the first embodiment of second aspect, described device further include:
Dictionary construction unit, for constructing theme dictionary relevant to current subject under discussion；
Word to formed unit, for by the theme dictionary each descriptor and emotion dictionary in each emotion word
Combination forms viewpoint word pair；
Second acquisition unit, for obtaining social media information relevant to current subject under discussion；
Third computing unit, for the similarity and each viewpoint according to each viewpoint word pair and every social media information
The viewpoint value of word pair and every social media information of similarity calculation of the comment of every social media information；
Information filtering unit, the social media information for being less than preset threshold for filtering viewpoint value, by remaining social matchmaker
Body information is as the information to be assessed.
In conjunction with the first embodiment of second aspect, in second of embodiment of second aspect, the dictionary structure
Unit is built to specifically include:
Subelement is searched for, for searching for social media information relevant to current subject under discussion in social networks；
Subelement is counted, for extracting the keyword in the social media information and counting the frequency that each keyword occurs
Rate；
Dictionary constructs subelement, for choosing the keyword of preset quantity as theme according to the sequence of frequency from high to low
Word constructs the theme dictionary.
In conjunction with the first embodiment of second aspect, in the third embodiment of second aspect, the third meter
Unit is calculated to be specifically used for:
The descriptor and the similarity of each keyword in a social media information for calculating a viewpoint word centering, mention
Take similarity maximum value a；Calculate each pass in the descriptor of the viewpoint word centering and the comment of the social media information
The similarity of keyword extracts similarity maximum value x；
The emotion word of the viewpoint word centering and the similarity of each emotion word in the social media information are calculated, is mentioned
Take similarity maximum value b；Calculate each feelings in the emotion word of the viewpoint word centering and the comment of the social media information
Feel the similarity of word, extracts similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 less than 1, institute
The similarity for stating viewpoint word pair and the comment of the social media information is s2=μ x+ (1- μ) y, and μ is greater than 0 less than 1；
By the viewpoint word pair and the similarity of the social media information and the viewpoint word pair and the social media
The similarity of the comment of information is added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained into all viewpoint subvalues of the social media information to same processing is done, by all sights
Point subvalue is cumulative to obtain the viewpoint value of the social media information, and so on, obtain the viewpoint value of each social media information.
In conjunction with the first of second aspect or second aspect embodiment or second of embodiment of second aspect, or
The third embodiment of second aspect, in the 4th kind of embodiment of second aspect, first computing unit is specifically used
In:
Determine the classification for the uncertain content for including in every information to be assessed；
Calculate the category score that the every class for including in every information to be assessed does not know content；
The category score that the every class for including in every information to be assessed does not know content is added up multiplied by after preset weight
Obtain the uncertain score of every information to be assessed.
In conjunction with the first of second aspect or second aspect embodiment or second of embodiment of second aspect, or
The third embodiment of second aspect, in the 5th kind of embodiment of second aspect, the reliability assessment unit is being incited somebody to action
The uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every information to be assessed
During supporting that ratio shared by opinion is inputted and calculated in trained quantitative appraisement model in advance in comment, it is described to
The uncertain score for assessing information is higher, and the confidence level of the information to be assessed is lower；The publisher of the information to be assessed
Confidence level it is lower, the confidence level of the information to be assessed is lower；It is supported shared by opinion in the comment of the information to be assessed
Ratio is smaller, and/or with the variation of time, supports ratio shared by opinion increasingly in the comment of the information to be assessed
Small, the confidence level of the information to be assessed is lower.
In the embodiment of the present invention, by the uncertain score of every information to be assessed, the publisher of every information to be assessed
Confidence level and every information to be assessed comment in support opinion shared by ratio input in advance trained quantitative evaluation mould
It is calculated in type, every information to be assessed is assessed by these three types of data, increases the accuracy of assessment.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, needed in being described below to the embodiment of the present invention
Attached drawing to be used is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention,
For those skilled in the art, other attached drawings can also be obtained such as these attached drawings.
Fig. 1 is that social media information credibility evaluation method one embodiment provided by the invention based on opining mining is shown
It is intended to；
Fig. 2 is another embodiment of the social media information credibility evaluation method provided by the invention based on opining mining
Schematic diagram；
Fig. 3 is one embodiment of the social media information reliability assessment device provided by the invention based on opining mining
Schematic diagram；
Fig. 4 is another embodiment of the social media information reliability assessment device provided by the invention based on opining mining
Schematic diagram；
Fig. 5 is another embodiment of the social media information reliability assessment device provided by the invention based on opining mining
Schematic diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those skilled in the art's every other embodiment obtained, shall fall within the protection scope of the present invention.
As shown in Figure 1, the embodiment of the present invention provides a kind of social media information reliability assessment side based on opining mining
Method, which comprises
101, information to be assessed is obtained；
Information to be assessed is the information extracted from the social media information in social networks, these information are all discussed with current
Topic is related, and social networks can be microblogging, wechat, Twitter etc..
102, the uncertain score of every information to be assessed is calculated；
This step is mainly used for judging whether contain uncertain content and the information in every information to be assessed not
Determine degree.
103, the confidence level of the publisher of every information to be assessed is calculated；
The calculating of the confidence level of information publisher is based primarily upon various features of the information publisher on social networks, example
Such as: delivering the number of microblogging, whether be certification user, user gradation etc. carries out reliability assessment, specific appraisal procedure to user
It can refer to existing method, details are not described herein again.
104, it counts in the comment of every information to be assessed and supports ratio shared by opinion；
105, by the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every
Support ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of item information to be assessed,
The output of quantitative appraisement model is the reliability order of every information to be assessed.
It should be noted that above-mentioned steps 102 to 104 are in the concrete realization, execution sequence, can be with there is no successive point
It executes side by side.
In the present embodiment, by the uncertain score of every information to be assessed, the publisher's of every information to be assessed can
Ratio shared by opinion is supported to input in trained quantitative appraisement model in advance in the comment of reliability and every information to be assessed
It is calculated, every information to be assessed is assessed by these three types of data, increases the accuracy of assessment.
For ease of understanding, information credibility appraisal procedure of the present invention is described with a specific embodiment below, is asked
Referring to Fig.2, the method for the present embodiment includes:
201, theme dictionary relevant to current subject under discussion is constructed；
In the present embodiment, theme dictionary relevant to current subject under discussion can be constructed by Word-frequency, the specific method is as follows:
Social media information relevant to current subject under discussion is searched in social networks, extracts keyword and statistics in social media information
The frequency that each keyword occurs is constructed according to the keyword that the sequence of frequency from high to low chooses preset quantity as descriptor
Theme dictionary.
In a specific example, such as Huawei Company has issued p7 mobile phone, emerged quickly in social networks with
The relevant social media information of p7 mobile phone, it can search social media information relevant to current subject under discussion p7 mobile phone, extraction are searched
Rope to these information in keyword for example: Huawei, screen, Hai Si, millet etc. count the frequency that each keyword occurs,
Then the keyword for choosing the higher preset quantity of the frequency of occurrences constructs theme dictionary as descriptor.
In addition, in other examples, common potential topic model latent topic model can also be used
Construct theme dictionary relevant to current subject under discussion.
202, each descriptor in the theme dictionary and each emotion word combination in emotion dictionary are formed into viewpoint
Word pair；
Emotion dictionary can use existing mainstream sentiment dictionary, and each viewpoint word is to by a descriptor and an emotion
Word composition, viewpoint word to such as<appearance, it is beautiful>,<Hai Si, it is proud>.
203, social media information relevant to current subject under discussion is obtained；
In the specific implementation, can scan for the keyword in current subject under discussion as input and climb in social media
It takes.
204, according to each viewpoint word pair and the similarity of every social media information and each viewpoint word pair and every society
Hand over the viewpoint value of every social media information of similarity calculation of the comment of media information；
The similarity of any two word A, BWherein A0、B0The term vector of word A, B are respectively indicated, | | A0||
Indicate A0Norm, | | B0| | indicate B0Norm.
The descriptor of a viewpoint word centering is calculated first, in accordance with above-mentioned calculating formula of similarity and a social media is believed
The similarity of each keyword in breath extracts similarity maximum value a；Descriptor and the society of the viewpoint word centering are calculated simultaneously
The similarity of each keyword in the comment of media information is handed over, similarity maximum value x is extracted；
Next the emotion word and the social media information of the viewpoint word centering are calculated according to above-mentioned calculating formula of similarity
In each emotion word similarity, extract similarity maximum value b；The emotion word and the social activity of the viewpoint word centering are calculated simultaneously
The similarity of each emotion word in the comment of media information extracts similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 can be pre- less than 1, λ
If the viewpoint word pair and the similarity of the comment of the social media information are s2=μ x+ (1- μ) y, μ is greater than 0 can be pre- less than 1, μ
If.
The viewpoint word pair and the similarity of the social media information and the viewpoint word are commented with the social media information
The similarity of opinion is added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained into all viewpoint subvalues of the social media information to same processing is done, by all viewpoints
Subvalue is cumulative to obtain the viewpoint value of the social media information, and so on, obtain the viewpoint value of each social media information.
205, filtering viewpoint value is less than the social media information of preset threshold, using remaining social media information as described in
Information to be assessed；
In the present embodiment, it is believed that viewpoint value is less than the social media information of preset threshold without subjectively, clearly
Some viewpoints are expressed, such as only states something with having no emotional color or describes some product, this part social media
Information will be filtered；It is considered that viewpoint value be greater than or equal to preset threshold social media information subjectively, clear earth's surface
Some viewpoints have been reached, this part social media information often becomes hot spot of public opinions, cognition of the people to event or product is influenced,
Therefore the confidence level of this partial information is mainly assessed using this part social media information as information to be assessed in the present embodiment.
206, the uncertain score of every information to be assessed is calculated；
In the present embodiment, an information uncertainty assessment models can be first trained, it is uncertain interior by include in information
Appearance is classified, such as the uncertain content for including in information can be done following classification:
| Type | Clue word or phrase | Example sentence |
| Problem type | Really | Do you think chip in the really sea of p7? |
| Hear type | It is said that | It is said that p7 is on sale in Europe |
| Wish type | Really think | Really think just there is platform p7 now |
| Conviction type | It is believed that | I believe that I has platform p7 some day |
| Conditional | If | If salary raise, I can buy p7 |
| Possible type | It should | I should be able to buy p7 mobile phone |
In the specific implementation, can be to be assessed with determination every by clue word or the information to be assessed of phrase on-line checking every
Then the classification for the uncertain content for including in information calculates the class that the every class for including in every information to be assessed does not know content
Other score, the category score that the every class for including in every information to be assessed is not known to finally content tire out multiplied by after preset weight
Add to obtain the uncertain score of every information to be assessed.
For example, giving an information to be assessed, it may belong to multiple classifications simultaneously, for example belong to A, B, C tri- simultaneously
Class has an a possibility that score, the higher expression of score belongs to the category bigger, such as calculates according to each classification of model
To the uncertainty of the information to be assessed, to assign to the score of these three classifications be respectively SA、SB、SC, then the final letter to be assessed
The uncertainty of breath is scored at H=WA*SA+WB*SB*WC*SC, wherein WA、WB、WCFor weight coefficient, the value of three weight coefficients
Can be different, such as one weight coefficient can be set for each classification in advance as needed, certain three weight coefficients can also be with
Take same value.
207, the confidence level of the publisher of every information to be assessed is calculated；
The calculating of the confidence level of information publisher is based primarily upon various features of the information publisher on social networks, example
Such as: delivering the number of microblogging, whether be certification user, user gradation etc. carries out reliability assessment, specific appraisal procedure to user
It can refer to existing method, details are not described herein again.
208, it counts in the comment of every information to be assessed and supports ratio shared by opinion；
209, by the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every
Support ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of item information to be assessed,
The output of quantitative appraisement model is the reliability order of every information to be assessed.
Specifically during calculating, the uncertain score of information to be assessed is higher, the confidence level of the information to be assessed
It is lower；The confidence level of the publisher of information to be assessed is lower, and the confidence level of the information to be assessed is lower；The comment of information to be assessed
Ratio shared by middle support opinion is smaller, and/or with the variation of time, supports shared by opinion in the comment of information to be assessed
Ratio is smaller and smaller, and the confidence level of the information to be assessed is lower.
It should be noted that above-mentioned steps 206 to 208 are in the concrete realization, execution sequence, can be with there is no successive point
It executes side by side.
In the present embodiment, after obtaining social media information relevant to current subject under discussion, constructed descriptor is utilized
Library and emotion dictionary calculate the similarity of emotion word pair and social media information and its comment information, thus extraction subjectively, it is bright
The social information for really expressing some viewpoints is assessed, during assessment, by the uncertain of every information to be assessed
Property score, ratio shared by opinion is supported in the comment of the confidence level of the publisher of every information to be assessed and every information to be assessed
Example is inputted in trained quantitative appraisement model in advance and is calculated, and is commented by these three types of data every information to be assessed
Estimate, increases the accuracy of assessment.
In practical applications, can accurately be understood according to assessment result user in social networks to the view of a certain event,
Demand of the hot spot or user of user's concern to a certain product, so as to accurately be that user does some information recommendations or to production
Product do some improvement, to promote user experience.
Information credibility provided in an embodiment of the present invention assessment device is described below, referring to Fig. 3, the present embodiment
Device 300 include:
First acquisition unit 301, for obtaining information to be assessed；
First computing unit 302, for calculating the uncertain score of every information to be assessed；
Second computing unit 303 calculates the confidence level of the publisher of every information to be assessed；
Statistic unit 304 supports ratio shared by opinion in the comment for counting every information to be assessed；
Reliability assessment unit 305, for by the uncertain score of every information to be assessed, every information to be assessed
Ratio shared by opinion is supported to input trained quantization in advance in the comment of the confidence level of publisher and every information to be assessed
It is calculated in assessment models, the output of quantitative appraisement model is the reliability order of every information to be assessed.
In the present embodiment, reliability assessment unit is by the uncertain score of every information to be assessed, every letter to be assessed
Support ratio shared by opinion as quantitative appraisement model in the comment of the confidence level of the publisher of breath and every information to be assessed
Input calculated, every information to be assessed is assessed by these three types of data, increases the accuracy of assessment.
For ease of understanding, information credibility of the present invention assessment device is described with a specific embodiment below, is asked
Refering to Fig. 4, the device 400 of the present embodiment includes:
Dictionary construction unit 401, for constructing theme dictionary relevant to current subject under discussion；
Word to formed unit 402, for by the theme dictionary each descriptor and emotion dictionary in each feelings
Feel word combination and forms viewpoint word pair；
Second acquisition unit 403, for obtaining social media information relevant to current subject under discussion；
Third computing unit 404, for according to the similarity of each viewpoint word pair and every social media information and each
The viewpoint value of viewpoint word pair and every social media information of similarity calculation of the comment of every social media information；
Information filtering unit 405, the social media information for being less than preset threshold for filtering viewpoint value, by remaining social activity
Media information is as the information to be assessed；
First acquisition unit 406, for obtaining information to be assessed；
First computing unit 407, for calculating the uncertain score of every information to be assessed；
Second computing unit 408 calculates the confidence level of the publisher of every information to be assessed；
Statistic unit 409 supports ratio shared by opinion in the comment for counting every information to be assessed；
Reliability assessment unit 410, for by the uncertain score of every information to be assessed, every information to be assessed
Ratio shared by opinion is supported to input trained quantization in advance in the comment of the confidence level of publisher and every information to be assessed
It is calculated in assessment models, the output of quantitative appraisement model is the reliability order of every information to be assessed.
In addition, dictionary construction unit 401 specifically includes search subelement 4011, statistics subelement 4012 and dictionary building
Unit 4013, in which:
Search subelement 4011 is used for, and social media information relevant to current subject under discussion is searched in social networks；
Statistics subelement 4012 is used for, and is extracted the keyword in the social media information and is counted each keyword appearance
Frequency；
Dictionary building subelement 4013 is used for, and the keyword conduct of preset quantity is chosen according to the sequence of frequency from high to low
Descriptor constructs the theme dictionary.
To further understand, device 400 is assessed to information credibility in the present embodiment with a practical application scene below
Interactive mode between interior each unit is described, specific as follows:
Firstly, dictionary construction unit 401 can construct theme dictionary relevant to current subject under discussion by Word-frequency.Specifically
Ground can search for social media information relevant to current subject under discussion by search subelement 4011 in social networks, then statistics
Unit 4012, which extracts the keyword in the social media information that search subelement 4011 searches and counts each keyword, to be occurred
Frequency, dictionary constructs subelement 4013 and chooses the keyword of preset quantity as descriptor according to frequency sequence from high to low
Construct theme dictionary.
In a specific example, such as Huawei Company has issued p7 mobile phone, emerged quickly in social networks with
The relevant social media information of p7 mobile phone, search subelement 4011 can search for social matchmaker relevant to current subject under discussion p7 mobile phone
Body information, statistics subelement 4012 extract the keyword in these information searched for example: Huawei, screen, Hai Si, millet
Deng counting the frequency that each keyword occurs, then dictionary building subelement 4013 chooses the higher preset quantity of the frequency of occurrences
Keyword as descriptor construct theme dictionary.
In addition, in other examples, dictionary construction unit 401 can also use common potential topic model
Latent topic model constructs theme dictionary relevant to current subject under discussion.
Word is to each descriptor and emotion word formed in the theme dictionary that unit 402 constructs dictionary construction unit 401
Each emotion word combination in library forms viewpoint word pair, and emotion dictionary can use existing mainstream sentiment dictionary, each viewpoint
Word is formed to by a descriptor and an emotion word, viewpoint word to such as<appearance, it is beautiful>,<Hai Si, it is proud>.
Second acquisition unit 403 obtains social media information relevant to current subject under discussion, in the specific implementation, second obtains list
Member 403 can scan for and crawl in social media using the keyword in current subject under discussion as input related to current subject under discussion
Social media information.
The similarity of any two word A, BWherein A0、B0The term vector of word A, B are respectively indicated, | | A0||
Indicate the norm of A0, | | B0| | indicate B0Norm.
Similarity and each viewpoint word of the third computing unit 404 according to each viewpoint word pair with every social media information
To the viewpoint value of every social media information of similarity calculation of the comment with every social media information.
Specifically, third computing unit 404 can calculate a viewpoint word centering according to above-mentioned calculating formula of similarity
The similarity of descriptor and each keyword in a social media information, extracts similarity maximum value a；The sight is calculated simultaneously
The similarity of the descriptor of point word centering and each keyword in the comment of the social media information, extracts similarity maximum value
x；
Following third computing unit 404 according to above-mentioned calculating formula of similarity calculate the emotion word of the viewpoint word centering with
The similarity of each emotion word in the social media information extracts similarity maximum value b；The viewpoint word centering is calculated simultaneously
The similarity of each emotion word in the comment of emotion word and the social media information extracts similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 can be pre- less than 1, λ
If the viewpoint word pair and the similarity of the comment of the social media information are s2=μ x+ (1- μ) y, μ is greater than 0 can be pre- less than 1, μ
If.
Third computing unit 404 is by the viewpoint word pair and the similarity of the social media information and the viewpoint word pair and the society
The similarity of the comment of media information is handed over to be added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained all viewpoints of the social media information to same processing is done by third computing unit 404
All viewpoint subvalues are added up and obtain the viewpoint value of the social media information by subvalue, and so on, obtain each social media letter
The viewpoint value of breath.
Information filtering unit 405 filters the social media information that viewpoint value is less than preset threshold, by remaining social media
For information as the information to be assessed, first acquisition unit 406 obtains remaining social media after information filtering unit 405 filters
Information.
In the present embodiment, it is believed that viewpoint value is less than the social media information of preset threshold without subjectively, clearly
Some viewpoints are expressed, such as only states something with having no emotional color or describes some product, this part social media
Information will be filtered；It is considered that viewpoint value be greater than or equal to preset threshold social media information subjectively, clear earth's surface
Some viewpoints have been reached, this part social media information often becomes hot spot of public opinions, cognition of the people to event or product is influenced,
Therefore the confidence level of this partial information is mainly assessed using this part social media information as information to be assessed in the present embodiment.
First computing unit 407 calculates the uncertain score of every information to be assessed, in the present embodiment, first calculate it is single
Member 407 can first train an information uncertainty assessment models, and the uncertain content for including in information is classified, such as
The uncertain content for including in information can be done following classification:
| Type | Clue word or phrase | Example sentence |
| Problem type | Really | Do you think chip in the really sea of p7? |
| Hear type | It is said that | It is said that p7 is on sale in Europe |
| Wish type | Really think | Really think just there is platform p7 now |
| Conviction type | It is believed that | I believe that I has platform p7 some day |
| Conditional | If | If salary raise, I can buy p7 |
| Possible type | It should | I should be able to buy p7 mobile phone |
In the specific implementation, the first computing unit 407 can by clue word or the information to be assessed of phrase on-line checking every, with
Then the classification for determining the uncertain content for including in every information to be assessed calculates the every class for including in every information to be assessed
The category score of uncertain content, finally the every class for including in every information to be assessed is not known the category score of content multiplied by
It adds up after preset weight and obtains the uncertain score of every information to be assessed.
For example, giving an information to be assessed, it may belong to multiple classifications simultaneously, for example belong to A, B, C tri- simultaneously
Class has an a possibility that score, the higher expression of score belongs to the category bigger, such as calculates according to each classification of model
To the uncertainty of the information to be assessed, to assign to the score of these three classifications be respectively SA、SB、SC, then the first computing unit 407
The uncertainty for the final information to be assessed being calculated is scored at H=WA*SA+WB*SB*WC*SC, wherein WA、WB、WCFor power
The value of weight coefficient, three weight coefficients can be different, such as a weight system can be arranged for each classification in advance as needed
Number, certain three weight coefficients can also take same value.
Second computing unit 408 calculates the confidence level of the publisher of every information to be assessed, the confidence level of information publisher
Calculating be based primarily upon various features of the information publisher on social networks, such as: deliver microblogging number, whether be certification
User, user gradation etc. carry out reliability assessment to user, and specific appraisal procedure can refer to existing method, and details are not described herein again.
Statistic unit 409, which counts, supports ratio shared by opinion, reliability assessment list in the comment of every information to be assessed
Member 410 by the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every it is to be evaluated
Estimate and support ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of information, quantization is commented
The output for estimating model is the reliability order of every information to be assessed.It is to be evaluated during reliability assessment unit 410 calculates
The uncertain score for estimating information is higher, and the confidence level of the information to be assessed is lower；The confidence level of the publisher of information to be assessed
Lower, the confidence level of the information to be assessed is lower；Support ratio shared by opinion smaller in the comment of information to be assessed, and/or
With the variation of time, in the comment of information to be assessed support opinion shared by ratio it is smaller and smaller, the information to be assessed can
Reliability is lower.
In the present embodiment, after second acquisition unit obtains social media information relevant to current subject under discussion, third meter
Calculate unit using theme dictionary constructed by dictionary construction unit and emotion dictionary calculate emotion word pair and social media information and
The similarity of its comment information is assessed to extract and subjectively, clearly express the social informations of some viewpoints, can
During reliability assessment unit assesses every information to be assessed, by the uncertain score of every information to be assessed,
Ratio conduct shared by opinion is supported in the comment of the confidence level of the publisher of every information to be assessed and every information to be assessed
The input of quantitative appraisement model is calculated, and is assessed by these three types of data every information to be assessed, is increased assessment
Accuracy.
Below referring to Fig. 5, Fig. 5 provides information credibility assessment another embodiment schematic diagram of device of the present invention, this reality
The information credibility assessment device 500 for applying example can be used for implementing information credibility appraisal procedure provided by the above embodiment,
In practical application, information credibility assessment device 500 is desirably integrated into electronic equipment, which can be mobile phone, puts down
Plate apparatus such as computer.Specifically:
Information credibility assessment device 500 may include RF (Radio Frequency, radio frequency) circuit 510, include one
A or more than one computer readable storage medium memory 520, input unit 530, display unit 540, sensor 550,
Voicefrequency circuit 560, includes one or more than one place at WiFi (wireless fidelity, Wireless Fidelity) module 570
Manage the components such as processor 580 and the power supply 590 of core.It will be understood by those skilled in the art that structure shown in Fig. 5 is simultaneously
The restriction to information credibility assessment device 500 is not constituted, may include components more more or fewer than diagram, or combine certain
A little components or different component layouts.Wherein:
RF circuit 510 can be used in messaging or communication process, and signal sends and receivees, particularly, by base station
After downlink information receives, one or the processing of more than one processor 580 are transferred to；In addition, the data for being related to uplink are sent to
Base station.In general, RF circuit 510 includes but is not limited to antenna, at least one amplifier, tuner, one or more oscillators, uses
Family identity module (SIM) card, transceiver, coupler, LNA (Low Noise Amplifier, low-noise amplifier), duplex
Device etc..In addition, RF circuit 510 can also be communicated with network and other equipment by wireless communication.The wireless communication can make
With any communication standard or agreement, and including but not limited to GSM (Global System of Mobile communication, entirely
Ball mobile communcations system), GPRS (General Packet Radio Service, general packet radio service), CDMA (Code
Division Multiple Access, CDMA), WCDMA (Wideband Code Division Multiple
Access, wideband code division multiple access), LTE (Long Term Evolution, long term evolution), Email, SMS (Short
Messaging Service, short message service) etc..
Memory 520 can be used for storing software program and module, and processor 580 is stored in memory 520 by operation
Software program and module, thereby executing various function application and data processing.Memory 520 can mainly include storage journey
Sequence area and storage data area, wherein storing program area can the (ratio of application program needed for storage program area, at least one function
Such as sound-playing function, image player function) etc.；Storage data area, which can be stored, creates data according to the use of storage equipment
(such as audio data, phone directory etc.).In addition, memory 520 may include high-speed random access memory, it can also include non-
Volatile memory, for example, at least a disk memory, flush memory device or other volatile solid-state parts.Accordingly
Ground, memory 520 can also include Memory Controller, to provide processor 580 and input unit 530 to memory 520
Access.
Input unit 530 can be used for receiving the number or character information of input, and generate and user setting and function
Control related keyboard, mouse, operating stick, optics or trackball signal input.Specifically, input unit 530 may include touching
Sensitive surfaces 531 and other input equipments 532.Touch sensitive surface 531, also referred to as touch display screen or Trackpad are collected and are used
Family on it or nearby touch operation (such as user using any suitable object or attachment such as finger, stylus in touch-sensitive table
Operation on face 531 or near touch sensitive surface 531), and corresponding attachment device is driven according to preset formula.It is optional
, touch sensitive surface 531 may include both touch detecting apparatus and touch controller.Wherein, touch detecting apparatus detection is used
The touch orientation at family, and touch operation bring signal is detected, transmit a signal to touch controller；Touch controller is from touch
Touch information is received in detection device, and is converted into contact coordinate, then gives processor 580, and can receive processor 580
The order sent simultaneously is executed.Furthermore, it is possible to using multiple types such as resistance-type, condenser type, infrared ray and surface acoustic waves
Realize touch sensitive surface 531.In addition to touch sensitive surface 531, input unit 530 can also include other input equipments 532.Specifically,
Other input equipments 532 can include but is not limited to physical keyboard, function key (such as volume control button, switch key etc.),
One of trace ball, mouse, operating stick etc. are a variety of.
Display unit 540 can be used for showing information input by user or be supplied to user information and device it is various
Graphical user interface, these graphical user interface can be made of figure, text, icon, video and any combination thereof.Display
Unit 540 may include display panel 541, optionally, can use LCD (Liquid Crystal Display, liquid crystal display
Device), the forms such as OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) configure display panel 541.
Further, touch sensitive surface 531 can cover display panel 541, when touch sensitive surface 531 detects that touch on it or nearby is grasped
After work, processor 580 is sent to determine the type of touch event, is followed by subsequent processing device 580 according to the type of touch event aobvious
Show and corresponding visual output is provided on panel 541.Although touch sensitive surface 531 and display panel 541 are as two in Fig. 5
Independent component realizes input and input function, but in some embodiments it is possible to by touch sensitive surface 531 and display panel
541 integrate and realize and output and input function.
Information credibility assessment device 500 may also include at least one sensor 550, such as optical sensor, motion-sensing
Device and other sensors.Specifically, optical sensor may include ambient light sensor and proximity sensor, wherein environment light passes
Sensor can adjust the brightness of display panel 541 according to the light and shade of ambient light, and proximity sensor can be moved to ear in device 500
Bian Shi closes display panel 541 and/or backlight.As a kind of motion sensor, gravity accelerometer can detect each
The size of (generally three axis) acceleration, can detect that size and the direction of gravity, can be used to identify device on direction when static
The application (such as horizontal/vertical screen switching, dependent game, magnetometer pose calibrating) of posture, (such as the step counting of Vibration identification correlation function
Device, percussion) etc.；Gyroscope, barometer, hygrometer, thermometer, infrared sensor for can also configure as device 500 etc. its
His sensor, details are not described herein.
Voicefrequency circuit 560, loudspeaker 561, microphone 562 can provide the audio interface between user and device.Audio-frequency electric
Electric signal after the audio data received conversion can be transferred to loudspeaker 561, be converted to sound by loudspeaker 561 by road 560
Signal output；On the other hand, the voice signal of collection is converted to electric signal by microphone 562, is turned after being received by voicefrequency circuit 560
It is changed to audio data, then by after the processing of audio data output processor 580, such as another device is sent to through RF circuit 510,
Or audio data is exported to memory 520 to be further processed.Voicefrequency circuit 560 is also possible that earphone jack, with
The communication of peripheral hardware earphone and device is provided.
WiFi belongs to short range wireless transmission technology, and information credibility assessment device 500 can be helped by WiFi module 570
Help user to send and receive e-mail, browse webpage and access streaming video etc., it provides wireless broadband internet for user and visits
It asks.Although Fig. 5 shows WiFi module 570, but it is understood that, and it is not belonging to must be configured into for device, completely may be used
To omit within the scope of not changing the essence of the invention as needed.
Processor 580 is the control centre of information credibility assessment device, is entirely filled using various interfaces and connection
The various pieces set, by running or executing the software program and/or module that are stored in memory 520, and calling storage
Data in memory 520 execute the various functions and processing data of storage equipment, to carry out whole prison to storage equipment
Control.Optionally, processor 580 may include one or more processing cores；Preferably, processor 580 can integrate application processor
And modem processor, wherein the main processing operation system of application processor, user interface and application program etc., modulatedemodulate
Processor is adjusted mainly to handle wireless communication.It is understood that above-mentioned modem processor can not also be integrated into processor
In 580.
It further includes the power supply 590 (such as battery) powered to all parts that information credibility assessment device, which sets 500, preferably
, power supply can be logically contiguous by power-supply management system and processor 580, to realize that management is filled by power-supply management system
The functions such as electricity, electric discharge and power managed.Power supply 590 can also include one or more direct current or AC power source, again
The random components such as charging system, power failure detection circuit, power adapter or inverter, power supply status indicator.
Although being not shown, information credibility assesses device 500 can also be including camera, bluetooth module etc., herein no longer
It repeats.Specifically in the present embodiment, information credibility assessment device 500 include memory 520 and one or one with
On program, one of them perhaps more than one program be stored in memory 520 and be configured to by one or one with
Upper processor 580 executes said one or more than one program includes the instruction for performing the following operation:
Obtain information to be assessed；
Calculate the uncertain score of every information to be assessed；
Calculate the confidence level of the publisher of every information to be assessed；
It counts and supports ratio shared by opinion in the comment of every information to be assessed；
By the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every to
It assesses and supports ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of information, quantify
The output of assessment models is the reliability order of every information to be assessed.
It should be noted that information credibility provided in an embodiment of the present invention assesses device 500, can be also used in realization
Other functions in Installation practice are stated, details are not described herein.
In addition it should be noted that, the apparatus embodiments described above are merely exemplary, wherein described as separation
The unit of part description may or may not be physically separated, component shown as a unit can be or
It can not be physical unit, it can it is in one place, or may be distributed over multiple network units.It can be according to reality
Border needs to select some or all of the modules therein to achieve the purpose of the solution of this embodiment.In addition, provided by the invention
In Installation practice attached drawing, the connection relationship between module indicates there is communication connection between them, specifically can be implemented as one
Item or a plurality of communication bus or signal wire.Those of ordinary skill in the art are without creative efforts, it can
It understands and implements.
Through the above description of the embodiments, it is apparent to those skilled in the art that the present invention can borrow
Help software that the mode of required common hardware is added to realize, naturally it is also possible to by specialized hardware include specific integrated circuit, specially
It is realized with CPU, private memory, special components and parts etc..Under normal circumstances, all functions of being completed by computer program are ok
It is easily realized with corresponding hardware, moreover, being used to realize that the specific hardware structure of same function is also possible to a variety of more
Sample, such as analog circuit, digital circuit or special circuit etc..But software program is real in situations more for the purpose of the present invention
It is now more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words makes the prior art
The part of contribution can be embodied in the form of software products, which is stored in the storage medium that can be read
In, such as the floppy disk of computer, USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
Device (RAM, Random Access Memory), magnetic or disk etc., including some instructions are with so that a computer is set
Standby (can be personal computer, server or the network equipment etc.) executes method described in each embodiment of the present invention.
It is provided for the embodiments of the invention a kind of social media information reliability assessment side based on opining mining above
Method and device are described in detail, for those of ordinary skill in the art, thought according to an embodiment of the present invention, specific
There will be changes in embodiment and application range, and therefore, the contents of this specification are not to be construed as limiting the invention.
Claims (12)
1. a kind of social media information credibility evaluation method based on opining mining characterized by comprising
Obtain social media information relevant to current subject under discussion；
Believed with the similarity of every social media information and each viewpoint word pair with every social media according to each viewpoint word pair
The viewpoint value of every social media information of similarity calculation of the comment of breath, wherein each viewpoint word to by descriptor and
Emotion word composition, the descriptor are obtained from theme dictionary relevant to the current subject under discussion；
The social media information that viewpoint value is less than preset threshold is filtered, using remaining social media information as information to be assessed；
Information to be assessed is obtained, the information to be assessed is to influence user to the social media information of the cognition of event or product；
Calculate the uncertain score of every information to be assessed；
Calculate the confidence level of the publisher of every information to be assessed；
It counts and supports ratio shared by opinion in the comment of every information to be assessed；
By the uncertain score of every information to be assessed, the confidence level of the publisher of every information to be assessed and every it is to be assessed
Support ratio shared by opinion inputs in trained quantitative appraisement model in advance to be calculated in the comment of information, the quantization
The output of assessment models is the reliability order of every information to be assessed.
2. the method as described in claim 1, which is characterized in that before obtaining information to be assessed, the method also includes:
Building theme dictionary relevant to the current subject under discussion；
Each descriptor in the theme dictionary and each emotion word combination in emotion dictionary are formed into the viewpoint word pair.
3. method according to claim 2, which is characterized in that relevant to the current subject under discussion theme dictionary of the building specifically wraps
It includes:
Social media information relevant to current subject under discussion is searched in social networks；
It extracts the keyword in the social media information and counts the frequency that each keyword occurs；
The theme dictionary is constructed as descriptor according to the keyword that the sequence of frequency from high to low chooses preset quantity.
4. method according to claim 2, which is characterized in that described according to each viewpoint word pair and every social media information
Similarity and each viewpoint word pair and every social media information of similarity calculation of the comment of every social media information
Viewpoint value specifically includes:
The descriptor and the similarity of each keyword in a social media information for calculating a viewpoint word centering, extract phase
Like degree maximum value a；Calculate each keyword in the descriptor of the viewpoint word centering and the comment of the social media information
Similarity, extract similarity maximum value x；
The emotion word of the viewpoint word centering and the similarity of each emotion word in the social media information are calculated, phase is extracted
Like degree maximum value b；Calculate each emotion word in the emotion word of the viewpoint word centering and the comment of the social media information
Similarity, extract similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 less than 1, the sight
Point word pair and the similarity of the comment of the social media information are s2=μ x+ (1- μ) y, and μ is greater than 0 less than 1；
By the viewpoint word pair and the similarity of the social media information and the viewpoint word pair and the social media information
The similarity of comment be added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained into all viewpoint subvalues of the social media information to same processing is done, by all viewpoint
Value is cumulative to obtain the viewpoint value of the social media information, and so on, obtain the viewpoint value of each social media information.
5. the method as described in any one of Claims 1-4, which is characterized in that described every information to be assessed of calculating
Uncertain score includes:
Determine the classification for the uncertain content for including in every information to be assessed；
Calculate the category score that the every class for including in every information to be assessed does not know content；
The category score that the every class for including in every information to be assessed does not know content is obtained multiplied by adding up after preset weight
The uncertain score of every information to be assessed.
6. the method as described in any one of Claims 1-4, which is characterized in that by every information to be assessed not really
Qualitative score is supported shared by opinion in the comment of the confidence level of the publisher of every information to be assessed and every information to be assessed
During being calculated in the preparatory trained quantitative appraisement model of ratio input, the uncertain of the information to be assessed is obtained
Point higher, the confidence level of the information to be assessed is lower；The confidence level of the publisher of the information to be assessed is lower, described to be evaluated
The confidence level for estimating information is lower；Support ratio shared by opinion smaller in the comment of the information to be assessed, and/or with the time
Variation, support ratio shared by opinion smaller and smaller in the comment of the information to be assessed, the information to be assessed it is credible
It spends lower.
7. a kind of social media information reliability assessment device based on opining mining characterized by comprising
Second acquisition unit, for obtaining social media information relevant to current subject under discussion；
Third computing unit, for according to each viewpoint word pair with every social media information similarity and each viewpoint word pair
With the viewpoint value of every social media information of similarity calculation of the comment of every social media information, wherein each sight
To being made of descriptor and emotion word, the descriptor obtains point word from theme dictionary relevant to the current subject under discussion；
Information filtering unit, the social media information for being less than preset threshold for filtering viewpoint value, remaining social media is believed
Breath is used as information to be assessed；
First acquisition unit, for obtaining information to be assessed, the information to be assessed is to influence user to recognize event or product
The social media information known；
First computing unit, for calculating the uncertain score of every information to be assessed；
Second computing unit calculates the confidence level of the publisher of every information to be assessed；
Statistic unit supports ratio shared by opinion in the comment for counting every information to be assessed；
Reliability assessment unit, for by the uncertain score of every information to be assessed, the publisher of every information to be assessed
Confidence level and every information to be assessed comment in support opinion shared by ratio input in advance trained quantitative evaluation mould
It is calculated in type, the output of the quantitative appraisement model is the reliability order of every information to be assessed.
8. device as claimed in claim 7, which is characterized in that described device further include:
Dictionary construction unit, for constructing theme dictionary relevant to the current subject under discussion；
Word to formed unit, for by the theme dictionary each descriptor and emotion dictionary in each emotion word combination
Form the viewpoint word pair.
9. device as claimed in claim 8, which is characterized in that the dictionary construction unit specifically includes:
Subelement is searched for, for searching for social media information relevant to current subject under discussion in social networks；
Subelement is counted, for extracting the keyword in the social media information and counting the frequency that each keyword occurs；
Dictionary constructs subelement, for choosing the keyword of preset quantity as descriptor structure according to the sequence of frequency from high to low
Build the theme dictionary.
10. device as claimed in claim 8, which is characterized in that the third computing unit is specifically used for:
The descriptor and the similarity of each keyword in a social media information for calculating a viewpoint word centering, extract phase
Like degree maximum value a；Calculate each keyword in the descriptor of the viewpoint word centering and the comment of the social media information
Similarity, extract similarity maximum value x；
The emotion word of the viewpoint word centering and the similarity of each emotion word in the social media information are calculated, phase is extracted
Like degree maximum value b；Calculate each emotion word in the emotion word of the viewpoint word centering and the comment of the social media information
Similarity, extract similarity maximum value y；
The similarity of the viewpoint word pair and the social media information is s1=λ a+ (1- λ) b, and λ is greater than 0 less than 1, the sight
Point word pair and the similarity of the comment of the social media information are s2=μ x+ (1- μ) y, and μ is greater than 0 less than 1；
By the viewpoint word pair and the similarity of the social media information and the viewpoint word pair and the social media information
The similarity of comment be added to obtain the viewpoint subvalue of the social media information；
Each viewpoint word is obtained into all viewpoint subvalues of the social media information to same processing is done, by all viewpoint
Value is cumulative to obtain the viewpoint value of the social media information, and so on, obtain the viewpoint value of each social media information.
11. the device as described in claim 7 to 10 any one, which is characterized in that first computing unit is specifically used for:
Determine the classification for the uncertain content for including in every information to be assessed；
Calculate the category score that the every class for including in every information to be assessed does not know content；
The category score that the every class for including in every information to be assessed does not know content is obtained multiplied by adding up after preset weight
The uncertain score of every information to be assessed.
12. the device as described in claim 7 to 10 any one, which is characterized in that the reliability assessment unit is will be every
The uncertain score of item information to be assessed, the confidence level of the publisher of every information to be assessed and every information to be assessed are commented
It is supported in during being calculated in the preparatory trained quantitative appraisement model of the input of ratio shared by opinion, it is described to be evaluated
The uncertain score for estimating information is higher, and the confidence level of the information to be assessed is lower；The publisher's of the information to be assessed
Confidence level is lower, and the confidence level of the information to be assessed is lower；Ratio shared by opinion is supported in the comment of the information to be assessed
Example is smaller, and/or with the variation of time, supports ratio shared by opinion smaller and smaller in the comment of the information to be assessed,
The confidence level of the information to be assessed is lower.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410436605.5A CN105447036B (en) | 2014-08-29 | 2014-08-29 | A kind of social media information credibility evaluation method and device based on opining mining |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410436605.5A CN105447036B (en) | 2014-08-29 | 2014-08-29 | A kind of social media information credibility evaluation method and device based on opining mining |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105447036A CN105447036A (en) | 2016-03-30 |
| CN105447036B true CN105447036B (en) | 2019-08-16 |
Family
ID=55557224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410436605.5A Active CN105447036B (en) | 2014-08-29 | 2014-08-29 | A kind of social media information credibility evaluation method and device based on opining mining |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105447036B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105824805A (en) * | 2016-05-09 | 2016-08-03 | 腾讯科技（深圳）有限公司 | Identification method and device |
| CN106649433B (en) * | 2016-09-05 | 2020-08-11 | 东南大学 | Topic opinion strength calculation method based on opinion statement credibility |
| CN107741938A (en) * | 2016-10-13 | 2018-02-27 | 腾讯科技（深圳）有限公司 | A kind of network information recognition methods and device |
| WO2018068664A1 (en) | 2016-10-13 | 2018-04-19 | 腾讯科技（深圳）有限公司 | Network information identification method and device |
| CN106528813B (en) * | 2016-11-18 | 2018-12-11 | 腾讯科技（深圳）有限公司 | A kind of multimedia recommendation method and device |
| CN108074071B (en) * | 2016-11-18 | 2021-06-18 | 腾讯科技（深圳）有限公司 | Project data processing method and device |
| CN106776551B (en) * | 2016-12-06 | 2020-05-08 | 桂林电子科技大学 | Method for analyzing emotion viewpoints of English composition |
| JP6835978B2 (en) * | 2017-02-21 | 2021-02-24 | ソニー・インタラクティブエンタテインメント エルエルシー | How to determine the authenticity of news |
| CN106951408A (en) * | 2017-03-17 | 2017-07-14 | 国信优易数据有限公司 | A kind of data digging method |
| CN109299400A (en) * | 2018-09-06 | 2019-02-01 | 北京奇艺世纪科技有限公司 | A kind of viewpoint abstracting method, device and equipment |
| CN109508370B (en) * | 2018-09-28 | 2022-07-08 | 北京百度网讯科技有限公司 | Comment extraction method, comment extraction device and storage medium |
| CN110059190A (en) * | 2019-04-18 | 2019-07-26 | 东南大学 | A kind of user's real-time point of view detection method based on social media content and structure |
| CN111539562A (en) * | 2020-04-10 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Data evaluation method and system based on model |
| CN112000709B (en) * | 2020-07-17 | 2023-10-24 | 微梦创科网络科技（中国）有限公司 | Social media information total exposure batch mining method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404591A (en) * | 2008-11-14 | 2009-04-08 | 西安交通大学 | Self-adapting dynamic trust weight estimation method |
| CN101404572A (en) * | 2008-11-14 | 2009-04-08 | 西安交通大学 | Network node total trust degree estimation method based on feedback trust aggregation |
| CN101466098A (en) * | 2009-01-21 | 2009-06-24 | 中国人民解放军信息工程大学 | Method, device and communication system for evaluating network trust degree |
| WO2013082395A1 (en) * | 2011-12-01 | 2013-06-06 | Google Inc | Identifying recommended merchants |
| CN103390194A (en) * | 2012-05-07 | 2013-11-13 | 北京三星通信技术研究有限公司 | Method, device and system for predicating user intention and recommending suggestion |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9390243B2 (en) * | 2012-02-28 | 2016-07-12 | Disney Enterprises, Inc. | Dynamic trust score for evaluating ongoing online relationships |
-
2014
- 2014-08-29 CN CN201410436605.5A patent/CN105447036B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101404591A (en) * | 2008-11-14 | 2009-04-08 | 西安交通大学 | Self-adapting dynamic trust weight estimation method |
| CN101404572A (en) * | 2008-11-14 | 2009-04-08 | 西安交通大学 | Network node total trust degree estimation method based on feedback trust aggregation |
| CN101466098A (en) * | 2009-01-21 | 2009-06-24 | 中国人民解放军信息工程大学 | Method, device and communication system for evaluating network trust degree |
| WO2013082395A1 (en) * | 2011-12-01 | 2013-06-06 | Google Inc | Identifying recommended merchants |
| CN103390194A (en) * | 2012-05-07 | 2013-11-13 | 北京三星通信技术研究有限公司 | Method, device and system for predicating user intention and recommending suggestion |
Non-Patent Citations (2)
| Title |
|---|
| 基于模糊可信度的软件质量的度量研究;李亚平;《长江大学学报(自科版)》;20140205;第11卷(第4期);11-15 |
| 微博新闻事件信息可信度评价;高雅;《中国优秀硕士学位论文全文数据库 信息科技辑》;20130915;正文第14,62-68页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105447036A (en) | 2016-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105447036B (en) | A kind of social media information credibility evaluation method and device based on opining mining | |
| CN106453053B (en) | Group message display methods and device | |
| EP3113035B1 (en) | Method and apparatus for grouping contacts | |
| CN104679969B (en) | Prevent the method and device of customer churn | |
| WO2016019925A1 (en) | Search method, server and client | |
| CN104239535B (en) | A kind of method, server, terminal and system for word figure | |
| CN108541310B9 (en) | Method and device for displaying candidate words and graphical user interface | |
| EP3031213B1 (en) | Apparatus and method for providing conversation topic | |
| CN105335398A (en) | Service recommendation method and terminal | |
| US20150058416A1 (en) | Determining a community emotional response | |
| US20160065731A1 (en) | Electronic device and method for displaying call information thereof | |
| CN107908619A (en) | Processing method, device, terminal and computer-readable storage medium based on public sentiment monitoring | |
| CN104244032A (en) | Method and device for pushing multimedia data | |
| CN103530562A (en) | Method and device for identifying malicious websites | |
| CN104618223B (en) | A kind of management method of information recommendation, device and system | |
| CN106126570B (en) | Information service system | |
| WO2020257988A1 (en) | Method for identifying click user, and related product | |
| US10204164B2 (en) | Systems and methods for filtering microblogs | |
| CN109543014B (en) | Man-machine conversation method, device, terminal and server | |
| WO2021003673A1 (en) | Content pushing method and related product | |
| CN107885718B (en) | Semantic determination method and device | |
| CN105550316B (en) | The method for pushing and device of audio list | |
| CN105094872B (en) | A kind of method and apparatus showing web application | |
| CN106486119A (en) | A kind of method and apparatus of identification voice messaging | |
| CN103488720A (en) | Method, system and client for viewing data |
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 |