CN113177163B - Method, system and storage medium for social dynamic information sentiment analysis - Google Patents
Method, system and storage medium for social dynamic information sentiment analysis Download PDFInfo
- Publication number
- CN113177163B CN113177163B CN202110468792.5A CN202110468792A CN113177163B CN 113177163 B CN113177163 B CN 113177163B CN 202110468792 A CN202110468792 A CN 202110468792A CN 113177163 B CN113177163 B CN 113177163B
- Authority
- CN
- China
- Prior art keywords
- information
- emotional tendency
- picture
- video
- uniform resource
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5846—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using extracted text
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7844—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using original textual content or text extracted from visual content or transcript of audio data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Abstract
The invention discloses a method for social dynamic information emotion analysis, which comprises the following steps: step S1: acquiring original social dynamic information of a user, preprocessing the original social dynamic information of the user to obtain processed social dynamic information, wherein the original social dynamic information of the user comprises: at least one of text information, picture uniform resource locator information, and video uniform resource locator information; step S2: calculating the emotional tendency probability of the processed social dynamic information; step S3: and obtaining social dynamic information emotion classification according to the emotional tendency probability. The emotional tendency probability can be calculated for at least one of the text, the picture and the video information type in the social dynamic information, and then the emotional tendency analysis is carried out on the user. The embodiment of the disclosure also discloses a system and a storage medium for social dynamic information emotion analysis.
Description
Technical Field
The invention relates to the technical field of emotion analysis, in particular to a method, a system and a storage medium for social dynamic information emotion analysis.
Background
Mobile social contact is becoming an important way for public information sharing and communication due to its characteristics such as convenience and openness. As mobile social users increase in size, the amount of user posted data also appears to grow explosively. To understand social sentiment and help government and enterprise scientific decisions, emotional tendency analysis based on user release content is becoming a research hotspot in the academic and industrial circles.
In the process of implementing the embodiments of the present disclosure, it is found that at least the following problems exist in the related art: the conventional method for carrying out emotion analysis on social dynamic information only carries out emotion analysis on single text, single picture or social dynamic information of the text and the picture, so that the scope of emotion analysis is limited, and the accuracy of emotion analysis results is not high.
Disclosure of Invention
The embodiment of the disclosure provides a method, a system and a storage medium for social dynamic information sentiment analysis, and aims to solve the technical problems that in the prior art, sentiment analysis is only performed on single text or picture social dynamic information, so that the sentiment analysis range is limited and the sentiment analysis result is not high in accuracy.
In a first aspect, a method for social dynamic information sentiment analysis is provided, the method comprising: step S1: acquiring original social dynamic information of a user, and preprocessing the original social dynamic information of the user to obtain processed social dynamic information, wherein the original social dynamic information of the user comprises: at least one of text information, picture uniform resource locator information, and video uniform resource locator information; step S2: calculating the emotional tendency probability of the processed social dynamic information; step S3: and obtaining social dynamic information emotion classification according to the emotional tendency probability.
With reference to the first aspect, in a first possible implementation manner of the first aspect, the step S2 further includes: step S21: and performing text feature extraction on the text information and calculating text emotional tendency probability.
With reference to the first aspect or the first possible implementation manner of the first aspect, in a second possible implementation manner of the first aspect, the step S2 further includes: step S22: clustering the picture uniform resource locator information, and calculating picture emotional tendency probability corresponding to the clustered picture uniform resource locator information; and/or, step S23: and clustering the video uniform resource locator information, and calculating the video emotional tendency probability corresponding to the clustered video uniform resource locator information.
With reference to the second possible implementation manner of the first aspect, in a third possible implementation manner of the first aspect, the step S3 further includes: step S31: calculating the emotional tendency probability R of the social dynamic information according to the following formula, wherein R is alpha multiplied by A URL +β×B URL + C, obtaining said social dynamic information sentiment classification according to R (pos, neg, neu), where α and β are piecewise functions, A URL For the picture emotional tendency probability, B URL And C is the video emotional tendency probability, pos is the probability that the emotional tendency of the social dynamic information is positive, neg is the probability that the emotional tendency of the social dynamic information is negative, and neu is the probability that the emotional tendency of the social dynamic information is neutral.
With reference to the second possible implementation manner of the first aspect, in a fourth possible implementation manner of the first aspect, the step S22 further includes: step S221: obtaining picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the picture uniform resource locator information is clustered; the step S23 further includes: step S231: and obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the video uniform resource locator information is clustered.
With reference to the second possible implementation manner of the first aspect, in a fifth possible implementation manner of the first aspect, the step S221 further includes: step S2211: obtaining picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering cluster according to the text emotional tendency probability value mean value corresponding to each clustering cluster after the picture uniform resource locator information is clustered; the step S231 further includes: step S2311: and obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the cluster according to the text emotional tendency probability value mean value corresponding to each cluster after the video uniform resource locator information is clustered.
With reference to the third possible implementation manner of the first aspect, in a sixth possible implementation manner of the first aspect, the first node is according to R (pos, neg) i Neu) obtaining the social dynamic information sentiment classification, including: and carrying out emotion classification on the original social dynamic information of the user according to the emotional tendency corresponding to the maximum value of the emotional tendency probability in the R (pos, neg, neu).
With reference to the first aspect, in a seventh possible implementation manner of the first aspect, the step S1 further includes: step S11: after the original social dynamic information of the user is obtained, whether the original social dynamic information of the user contains text information, picture uniform resource locator information or video uniform resource locator information is marked, and the marked original social dynamic information of the user is preprocessed.
In a second aspect, a system for social dynamic message sentiment analysis is provided, the system comprising: the social dynamic information acquisition module is used for acquiring original social dynamic information of a user, preprocessing the original social dynamic information of the user and acquiring processed social dynamic information, wherein the original social dynamic information of the user comprises: at least one of text information, picture uniform resource locator information, and video uniform resource locator information; the emotional tendency probability calculation module is used for calculating the emotional tendency probability of the processed social dynamic information; and the emotion classification acquisition module is used for obtaining social dynamic information emotion classification according to the emotion tendency probability.
In a third aspect, a storage medium is provided, the storage medium storing a computer program comprising program instructions, which when executed by a processor, cause the processor to perform the aforementioned method for social dynamics information sentiment analysis.
The method, the system and the storage medium for social dynamic information emotion analysis provided by the embodiment of the disclosure can achieve the following technical effects:
firstly, calculating emotional tendency probability of at least one of text, picture and video information types in social dynamic information, and further analyzing emotional tendency of a user; in addition, the emotional tendency analysis is carried out on the users according to the types of the social information of at least two users, the defect of single dimensionality of traditional emotional analysis data can be overcome, the comprehensiveness of the dimensionality of the emotional analysis is improved, and the accuracy of the emotional analysis result can be improved to a certain extent.
The foregoing general description and the following description are exemplary and explanatory only and are not restrictive of the application.
Drawings
One or more embodiments are illustrated by way of example in the accompanying drawings, which correspond to the figures, and not by way of limitation, in which elements having the same reference numeral designations are shown as similar elements and not to scale, and in which:
FIG. 1 is a schematic flow chart of a method for social dynamic information sentiment analysis provided by an embodiment of the present disclosure;
FIG. 2 is another schematic diagram of a flow of a method for social dynamic information sentiment analysis provided by an embodiment of the present disclosure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In the embodiments of the present disclosure, "and/or" describes an association relationship of associated objects, which means that there may be three relationships, for example, "a and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship.
In the process of implementing the embodiment of the disclosure, it is found that the existing emotion tendency analysis task is mainly developed around text data, picture and video information which can express network user emotion are ignored, and with the increasing growth of the word, picture and video information issued by the network platform user, a solution for emotion analysis of the word, picture and video information can be performed, so that the comprehensiveness and scientificity of emotion tendency analysis of the user can be further improved. In addition, the storage is difficult to meet the requirement due to the overlarge amount of the picture and video data, and great challenges are brought to the emotion tendency analysis task facing the picture and video data.
The social dynamics in the embodiments of the present disclosure may be microblog dynamics, wechat friend circle, QQ space dynamics, twitter or facebook dynamics, or other social application dynamics.
FIG. 1 is a flowchart of a method for social dynamic information sentiment analysis provided by an embodiment of the present disclosure. As shown in fig. 1, an embodiment of the present disclosure provides a method for social dynamic information sentiment analysis, the method including: step S1: acquiring original social dynamic information of a user, preprocessing the original social dynamic information of the user to obtain processed social dynamic information, wherein the original social dynamic information of the user comprises: at least one of text information, picture Uniform Resource Locator (URL) information and video URL; step S2: calculating the emotional tendency probability of the processed social dynamic information; step S3: and obtaining social dynamic information emotion classification according to the emotional tendency probability. The original social dynamic information of the user specifically comprises: text information, picture URL, video URL, text information and picture URL, text information and video URL, picture URL and video URL, or text information and picture URL and video URL.
The method for social dynamic information emotion analysis provided by the embodiment of the disclosure can achieve the following technical effects: firstly, calculating emotional tendency probability of at least one of text, picture and video information types in social dynamic information, and further analyzing emotional tendency of a user; in addition, emotional tendency analysis is performed on the user according to at least two types of user social information, the defect of single dimensionality of traditional emotional analysis data can be overcome, comprehensiveness of the dimensionality of the emotional analysis is improved, and accuracy of an emotional analysis result can be improved to a certain extent.
In some embodiments, step S2 further includes: step S21: and performing text feature extraction on the text information and calculating text emotional tendency probability. The method can utilize part of user social dynamic information marked with emotional tendency to finely adjust (also called fine-tuning) a Bidirectional coding Representation from transducers (BERT) model based on a converter to obtain a trained model, and extract emotional tendency information of text information by using the trained model. It should be noted that, those skilled in the art may also use albert (a Lite BERT), or roberta (robustly optimized BERT approach), as long as feature extraction can be performed on text information. The text information can be better subjected to feature extraction by using the BERT model.
In some embodiments, for the calculation of text emotion tendency probabilities, a SoftMax regression algorithm may be used. The emotional tendency classification can use two-classification, three-classification, six-classification or other emotional classification modes. The technical scheme involved in the embodiment of the disclosure is explained only in a three-classification emotion classification mode, and the text emotion tendency probability C ═ in a piece of social dynamic information is calculated and obtained through a SoftMax regression algorithm pos ,text_ neg ,text_ neu ) Wherein text u pos Text u for the probability that the emotional tendency of the text is positive neg Text u for the probability that the emotional tendency of the text is negative neu Is the probability that the emotional tendency of the text is neutral. In this way, the emotional tendency probability of the social dynamics of the user social information including the text information can be obtained.
In some embodiments, step S2 further includes: step S22: clustering the picture URL information, and calculating picture emotional tendency probability corresponding to the clustered picture URL information; and/or, step S23: and clustering the video URL information, and calculating the video emotional tendency probability corresponding to the clustered video URL information. Wherein, the step S2 may include: step S21, step S22, step S23, step S21, step S22, step S21, step S23, step S22, and step S23, or step S21, step S22, and step S23. Therefore, the emotional tendency of the user can be analyzed on the social dynamics of the user, wherein the social dynamics information comprises at least one of text information, picture URL information or video URL information, the emotional tendency analysis range of the user is expanded, and the emotional tendency analysis accuracy of the user is improved to a certain extent.
In some embodiments, the picture URL or the video URL is subjected to cluster analysis, and the picture URL and the video URL may be clustered respectively by using Affinity propagation Clustering Algorithm (AP Clustering Algorithm). So that picture or video URLs of the same or similar URLs are grouped into one category. Other clustering algorithms can be used for clustering the picture URLs or the video URLs, and different clustering algorithms can be used for clustering the picture URLs and the video URLs. The same pictures or videos published by the users of the social platform have the same URL, and similar pictures or videos published by the users of the social platform have similar URLs and can be obtained through network access. Therefore, clustering pictures or video URLs of the same or similar URLs can reduce the calculation amount of emotional tendency probability calculation on the picture URLs and the video URLs to a certain extent, and reduce the complexity of the calculation process; in addition, in the prior art, the pictures and videos are analyzed, and the method for calculating the emotional orientation cannot download corresponding data due to the fact that the picture and video data resources are deleted and cannot be accessed, and further analysis cannot be performed.
In some embodiments, step S22 further includes: step S221: obtaining picture emotional tendency probability corresponding to each picture URL information in the cluster according to the text emotional tendency probability value corresponding to each cluster after the picture URL information is clustered; step S23 further includes: step S231: and obtaining the video emotional tendency probability corresponding to the video URL information in each cluster according to the text emotional tendency probability value in each cluster after clustering. The image emotional tendency probability corresponding to each image URL information in the cluster can be obtained by removing the mean value of the maximum value, the minimum value or the small probability part according to the text emotional tendency probability value corresponding to each cluster after the image URL information is clustered, or selecting the text emotional tendency probability which can represent the average level or the overall level of a group of data, such as mode. And similarly, the video emotional tendency probability can be obtained. Therefore, the pictures and the videos are stored without consuming mass storage resources, the downloading and analyzing processes of the pictures and the videos can be avoided, the consumption of hardware resources required by the storage of the pictures and the videos is saved, a large amount of calculation and analysis on the pictures and the videos are not required to be performed by using a deep learning algorithm, the calculation efficiency is improved to a certain extent, and the complexity of analyzing the pictures and the videos is reduced.
In some embodiments, step S221 further comprises: step S2211: obtaining picture emotional tendency probability corresponding to each picture URL information in the cluster according to the text emotional tendency probability value mean value corresponding to each cluster after the picture URL information is clustered; alternatively, step S231 further includes: step S2311: and obtaining the video emotional tendency probability corresponding to each video URL information in the cluster according to the text emotional tendency probability value mean value corresponding to each cluster after the video URL information is clustered. The following description is given by taking the calculation of the emotional tendency probability of the picture corresponding to the picture URL information as an example, and it is assumed that the cluster set where the picture URL of the current emotional tendency probability value to be calculated is located is A, the set A includes n pieces of microblog texts, wherein the emotional tendency probability of the ith social dynamic text is
The emotional tendency probability value of the picture URL in the set A can be calculated by the following formula:
where i is 1, 2, … …, n. Similarly, the emotional tendency probability B of the video represented by the video URL information in the set A can be obtained URL . Therefore, the emotional tendency probability values of the pictures and the videos can be obtained by clustering the picture URLs and the video URLs and utilizing the text emotional tendency probability value means in various clusters after clustering, and the analysis of the picture and the video resources can be applied to the emotional tendency analysis of the social dynamic information of the users by using few computing resources.
In some embodiments, step S3 further includes: step S31: calculating the emotional tendency probability R ═ alpha multiplied by A of certain social dynamic information by the following formula URL +β×B URL + C, obtaining the social dynamic information sentiment classification according to R (pos, neg, neu), wherein alpha and beta are piecewise functions, A URL Picture emotional tendency probability for the piece of social dynamic information, B URL The video emotional tendency probability of the social dynamic information, C, pos, neg and neu are respectively the video emotional tendency probability of the social dynamic information, the text emotional tendency probability of the social dynamic information, the probability that the emotional tendency of the social dynamic information is positive, the probability that the emotional tendency of the social dynamic information is negative and the probability that the emotional tendency of the social dynamic information is neutral. If the social dynamic information includes text information, C ═ text ═ u can be obtained according to step S21 pos ，text_ neg ,text_ neu ) (ii) a Otherwise, C is 0. If the social dynamic information includes the picture information, then a can be obtained from the picture URL information according to step S2211 uRL (ii) a Otherwise, A URL 0. If the social dynamic information includes the picture information, then step S2311 is performed to obtain B from the picture URL information URL (ii) a Otherwise, B URL 0. The values of alpha and beta are obtained according to experience, and can be adjusted by combining with an actual service scene. For example, the number of URL's num in the cluster set A where the picture or video is located<When 10, α is 0.01; number of URL pieces num in cluster set A>When equal to 10 times and num<At 100, a is 0.05; when the number of URL is largenum>When 100, α is 0.2. The value principle of beta and alpha is consistent in the same way:
in this way, the emotional tendency of the user can be analyzed by fusing at least one of text information, picture information or video information of the social dynamics.
In some embodiments, obtaining the social dynamics information sentiment classification from R (pos, neg, neu) includes: and carrying out emotion classification on the original social dynamic information of the user according to the emotional tendency corresponding to the maximum value of the emotional tendency probability in the R (pos, neg, neu). In pos, neg and neu, if pos is the maximum value, the social dynamic information of the user belongs to the positive emotional tendency; if neg is the maximum value in pos, neg and neu, the social dynamic information of the user belongs to negative emotional tendency; if the neu is the maximum value in pos, neg and neu, the social dynamic information of the user belongs to neutral emotional tendency.
In some embodiments, step S1 further includes: step S11: after the original social dynamic information of the user is obtained, whether the original social dynamic information of the user contains text information, picture URL information or video URL information is marked, and the marked original social dynamic information of the user is preprocessed. Wherein the pretreatment comprises: data duplication removal, data cleaning, special character messy codes removal and the like, and the text, the picture URL and the video URL of the same user social dynamic information are associated together according to the collected fields.
FIG. 2 is another schematic diagram of a flow of a method for social dynamic information sentiment analysis provided by an embodiment of the present disclosure. As shown in fig. 2, step P1: after collecting the text, picture URL and video URL information in the user social application program, the process proceeds to step P2: after obtaining the text emotional tendency probability, the step proceeds to step P3: after obtaining the picture emotional tendency probability and/or the video emotional tendency probability, the method proceeds to step P4: and (3) finely adjusting the text emotional tendency probability, the picture emotional tendency probability and/or the video emotional tendency probability by combining the weight, and then turning to a step P5: after the fused social dynamic information emotional tendency probability is obtained, the step is switched to step P6: and obtaining social dynamic information emotion classification according to the social dynamic information emotion tendency probability.
Step P2 further includes: step P21: judging whether text information exists, if not, turning to the step P22: if the text message is marked as null, the process proceeds to step P23: and (4) performing data cleaning, and then switching to the step P24: after extracting the text information, the process proceeds to step P25: after model training and model parameter adjustment are performed on the BERT model, the method proceeds to step P26: after obtaining the trained BERT model, go to step P27: after extracting the text features of the text message by using the BERT model, the process proceeds to step P28: after the text features are mapped, the step P29 is executed: carrying out text emotional tendency probability calculation, and turning to a step P36; step P1 further includes: step P31: if yes, the process goes to step P33: if the URL is marked as null, the method proceeds to step P32: judging the URL category, wherein the step P32 comprises the following steps: step P34: picture URL, and, step P35: after the steps P34 and P35 are completed, the process proceeds to step P36: after associating the text, picture URL and video URL information in the social dynamic information of the same user, the process proceeds to step P37: after the picture URL and the video URL are subjected to clustering analysis, the step P38 is carried out: and (4) according to the text emotional tendency probability fusion in each clustered cluster after clustering, after obtaining the emotional tendency probability corresponding to the URL of the category, turning to the step P39: and obtaining the picture URL emotional tendency probability and/or the video URL emotional tendency probability.
The embodiment of the present disclosure further provides a system for social dynamic message emotion analysis, where the system includes: the social dynamic information acquisition module is used for acquiring original social dynamic information of the user, preprocessing the original social dynamic information of the user and obtaining processed social dynamic information, wherein the original social dynamic information of the user comprises: at least one of text information, picture URL information, and video URL information; the emotional tendency probability calculation module is used for calculating the emotional tendency probability of the processed social dynamic information; and the emotion classification acquisition module is used for obtaining the social dynamic information emotion classification according to the emotion tendency probability.
In some embodiments, the emotional tendency probability calculation module comprises: and the text feature extraction module is used for extracting text features of the text information and calculating the text emotional tendency probability.
In some embodiments, the emotional tendency probability calculation module further comprises: the picture clustering analysis module is used for clustering the picture uniform resource locator information and calculating picture emotional tendency probability corresponding to the clustered picture uniform resource locator information; and/or the video clustering analysis module is used for clustering the video uniform resource locator information and calculating the video emotion tendency probability corresponding to the clustered video uniform resource locator information.
In some embodiments, the emotion classification acquisition module comprises: the emotional tendency probability calculation module is used for calculating social dynamic information emotional extreme value R through the following formula, wherein R is alpha multiplied by A URL +β×B URL + C, obtaining social dynamic information sentiment classification according to R (pos, neg, neu), wherein alpha and beta are piecewise functions, A URL As picture emotional tendency probability, B URL The probability of the video emotional tendency is C, the probability of the text emotional tendency, pos is the probability that the emotional tendency of the social dynamic information is positive, neg is the probability that the emotional tendency of the social dynamic information is negative, and neu is the probability that the emotional tendency of the social dynamic information is neutral.
In some embodiments, the picture cluster analysis module comprises: and the picture clustering analysis unit is used for obtaining the picture emotional tendency probability corresponding to the uniform resource locator information of the pictures in the clustering clusters according to the text emotional tendency probability value corresponding to each clustering cluster after the picture URL information is clustered. A video cluster analysis module comprising: and the video clustering analysis unit is used for obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the video URL information is clustered.
In some embodiments, the picture cluster analysis unit further includes: and the first unit is used for obtaining the picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering cluster according to the text emotional tendency probability value mean value corresponding to each clustering cluster after the picture URL information is clustered. The video clustering analysis unit further comprises: and the second unit is used for obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the clustering cluster according to the text emotional tendency probability value mean value corresponding to each clustering cluster after the video URL information is clustered.
In some embodiments, the obtaining of the social dynamic information sentiment classification according to R (pos, neg, neu) in the sentiment classification obtaining module includes: and carrying out emotion classification on the original social dynamic information of the user according to the emotional tendency corresponding to the maximum value of the emotional tendency probability in the R (pos, neg, neu).
In some embodiments, the social dynamic information obtaining module comprises: the marking module is used for marking whether the original social dynamic information of the user contains text information, picture uniform resource locator information or video uniform resource locator information after the original social dynamic information of the user is obtained, and preprocessing the marked original social dynamic information of the user.
The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.
The disclosed embodiments also provide a storage medium storing a computer program comprising program instructions which, when executed by a processor, cause the processor to perform the aforementioned method for social dynamics information sentiment analysis.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (7)
1. A method for social dynamic information sentiment analysis, comprising:
step S1: acquiring original social dynamic information of a user, and preprocessing the original social dynamic information of the user to obtain processed social dynamic information, wherein the original social dynamic information of the user comprises: text information, picture uniform resource locator information and video uniform resource locator information;
step S2: calculating the emotional tendency probability of the processed social dynamic information;
step S21: extracting text features of the text information and calculating text emotional tendency probability;
step S22: clustering the picture uniform resource locator information, and calculating picture emotional tendency probability corresponding to the clustered picture uniform resource locator information;
step S23: clustering the video uniform resource locator information, and calculating video emotional tendency probability corresponding to the clustered video uniform resource locator information;
the step S22 further includes: step S221: obtaining picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the picture uniform resource locator information is clustered;
the step S23 further includes: step S231: obtaining video emotional tendency probability corresponding to the video uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the video uniform resource locator information is clustered;
step S3: and obtaining social dynamic information emotion classification according to the emotional tendency probability.
2. The method according to claim 1, wherein the step S3 further comprises:
step S31: calculating the probability R (pos, neg, neu) of emotional tendency of the social dynamic information through the following formula,
R＝α×A URL +β×B URL +C
obtaining the social dynamic information emotion classification according to R (pos, neg, neu), wherein alpha and beta are piecewise functions, A URL For the picture emotional tendency probability, B URL The video emotional tendency probability, the text emotional tendency probability, pos, neg and neu are respectively the probability that the emotional tendency of the social dynamic information is positive, the probability that the emotional tendency of the social dynamic information is negative and the probability that the emotional tendency of the social dynamic information is neutral;
wherein the content of the first and second substances,
num represents the number of URL's in the picture or video cluster set.
3. The method of claim 1,
the step S221 further includes: step S2211: obtaining picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering cluster according to the text emotional tendency probability value mean value corresponding to each clustering cluster after the picture uniform resource locator information is clustered;
the step S231 further includes: step S2311: and obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the cluster according to the text emotional tendency probability value mean value corresponding to each cluster after the video uniform resource locator information is clustered.
4. The method of claim 2, wherein obtaining the social dynamics information sentiment classification from R (pos, neg, neu) comprises: and carrying out emotion classification on the original social dynamic information of the user according to the emotional tendency corresponding to the maximum value of the emotional tendency probability in the R (pos, neg, neu).
5. The method according to claim 1, wherein the step S1 further comprises:
step S11: after the original social dynamic information of the user is obtained, whether the original social dynamic information of the user contains text information, picture uniform resource locator information or video uniform resource locator information is marked, and the marked original social dynamic information of the user is preprocessed.
6. A system for social dynamic message emotion analysis, comprising:
the social dynamic information acquisition module is used for acquiring original social dynamic information of a user, preprocessing the original social dynamic information of the user and acquiring processed social dynamic information, wherein the original social dynamic information of the user comprises: text information, picture uniform resource locator information and video uniform resource locator information;
the emotional tendency probability calculation module is used for calculating the emotional tendency probability of the processed social dynamic information;
the emotion classification acquisition module is used for obtaining social dynamic information emotion classification according to the emotion tendency probability;
the emotion tendency probability calculation module comprises a text feature extraction module and is used for extracting text features of the text information and calculating the text emotion tendency probability;
the emotion tendency probability calculation module also comprises a picture clustering analysis module which is used for clustering the picture uniform resource locator information and calculating the picture emotion tendency probability corresponding to the clustered picture uniform resource locator information;
the video clustering analysis module is used for clustering the video uniform resource locator information and calculating video emotional tendency probability corresponding to the clustered video uniform resource locator information;
the picture clustering analysis module comprises: the picture clustering analysis unit is used for obtaining the picture emotional tendency probability corresponding to the picture uniform resource locator information in the clustering clusters according to the text emotional tendency probability value corresponding to each clustering cluster after the picture uniform resource locator information is clustered;
a video cluster analysis module comprising: and the video clustering analysis unit is used for obtaining the video emotional tendency probability corresponding to the video uniform resource locator information in the clustering cluster according to the text emotional tendency probability value corresponding to each clustering cluster after the video uniform resource locator information is clustered.
7. A storage medium, characterized in that the storage medium stores a computer program comprising program instructions which, when executed by a processor, cause the processor to carry out the method for social dynamics information sentiment analysis of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110468792.5A CN113177163B (en) | 2021-04-28 | 2021-04-28 | Method, system and storage medium for social dynamic information sentiment analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110468792.5A CN113177163B (en) | 2021-04-28 | 2021-04-28 | Method, system and storage medium for social dynamic information sentiment analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113177163A CN113177163A (en) | 2021-07-27 |
| CN113177163B true CN113177163B (en) | 2022-08-02 |
Family
ID=76925266
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110468792.5A Active CN113177163B (en) | 2021-04-28 | 2021-04-28 | Method, system and storage medium for social dynamic information sentiment analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113177163B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764268A (en) * | 2018-04-02 | 2018-11-06 | 华南理工大学 | A kind of multi-modal emotion identification method of picture and text based on deep learning |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020712B (en) * | 2012-12-28 | 2015-10-28 | 东北大学 | A kind of distributed sorter of massive micro-blog data and method |
| CN103500175B (en) * | 2013-08-13 | 2017-09-15 | 中国人民解放军国防科学技术大学 | A kind of method based on sentiment analysis on-line checking microblog hot event |
| CN105117455A (en) * | 2015-08-18 | 2015-12-02 | 北京奇虎科技有限公司 | Along-the-road target image search method, terminal and system |
| CN105138991B (en) * | 2015-08-27 | 2016-08-31 | 山东工商学院 | A kind of video feeling recognition methods merged based on emotion significant characteristics |
| CN106649578A (en) * | 2016-11-17 | 2017-05-10 | 华北理工大学 | Public opinion analysis method and system based on social network platform |
| CN108334758B (en) * | 2017-01-20 | 2020-08-18 | 中国移动通信集团山西有限公司 | Method, device and equipment for detecting user unauthorized behavior |
| CN107341270B (en) * | 2017-07-28 | 2020-07-03 | 东北大学 | Social platform-oriented user emotion influence analysis method |
| CN108108849A (en) * | 2017-12-31 | 2018-06-01 | 厦门大学 | A kind of microblog emotional Forecasting Methodology based on Weakly supervised multi-modal deep learning |
| CN109508375A (en) * | 2018-11-19 | 2019-03-22 | 重庆邮电大学 | A kind of social affective classification method based on multi-modal fusion |
| CN109948148A (en) * | 2019-02-28 | 2019-06-28 | 北京学之途网络科技有限公司 | A kind of text information emotion determination method and decision maker |
| US11095588B2 (en) * | 2019-10-16 | 2021-08-17 | Accenture Global Solutions Limited | Social network data processing and profiling |
| CN111126194B (en) * | 2019-12-10 | 2023-04-07 | 郑州轻工业大学 | Social media visual content emotion classification method |
| CN111859925B (en) * | 2020-08-06 | 2023-08-08 | 东北大学 | Emotion analysis system and method based on probability emotion dictionary |
| CN112214601B (en) * | 2020-10-21 | 2022-06-10 | 厦门市美亚柏科信息股份有限公司 | Social short text sentiment classification method and device and storage medium |
-
2021
- 2021-04-28 CN CN202110468792.5A patent/CN113177163B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764268A (en) * | 2018-04-02 | 2018-11-06 | 华南理工大学 | A kind of multi-modal emotion identification method of picture and text based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 基于分类的微博情感分析算法研究及实现;杨艳霞;《计算机与数字工程》;20170220(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113177163A (en) | 2021-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105654950B (en) | Adaptive voice feedback method and device | |
| US11138903B2 (en) | Method, apparatus, device and system for sign language translation | |
| CN107609092B (en) | Intelligent response method and device | |
| CN104284306B (en) | A kind of method for filtering spam short messages, system, mobile terminal and Cloud Server | |
| CN111401063B (en) | Text processing method and device based on multi-pool network and related equipment | |
| CN117332072B (en) | Dialogue processing, voice abstract extraction and target dialogue model training method | |
| CN115759001A (en) | Language model training method, text prediction method and device | |
| CN111859210B (en) | Image processing method, device, equipment and storage medium | |
| CN111191141A (en) | Document recommendation method and device | |
| CN112801721B (en) | Information processing method, information processing device, electronic equipment and storage medium | |
| CN113177163B (en) | Method, system and storage medium for social dynamic information sentiment analysis | |
| CN111767953B (en) | Method and apparatus for training an article coding model | |
| CN110377706B (en) | Search sentence mining method and device based on deep learning | |
| CN111555960A (en) | Method for generating information | |
| CN116561639A (en) | Multi-mode data emotion analysis method for open source information | |
| CN110083654A (en) | A kind of multi-source data fusion method and system towards science and techniques of defence field | |
| CN113055890B (en) | Multi-device combination optimized real-time detection system for mobile malicious webpage | |
| CN115718904A (en) | Text processing method and device | |
| CN114118087A (en) | Entity determination method, entity determination device, electronic equipment and storage medium | |
| CN114782077A (en) | Information screening method, model training method, device, electronic equipment and medium | |
| CN114390452A (en) | Message sending method and related equipment | |
| CN113569067A (en) | Label classification method and device, electronic equipment and computer readable storage medium | |
| CN113486147A (en) | Text processing method and device, electronic equipment and computer readable medium | |
| CN112000813A (en) | Knowledge base construction method and device | |
| CN112748828A (en) | Information processing method, device, terminal equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |