CN111046137A - Multidimensional emotion tendency analysis method - Google Patents
Multidimensional emotion tendency analysis method Download PDFInfo
- Publication number
- CN111046137A CN111046137A CN201911108881.8A CN201911108881A CN111046137A CN 111046137 A CN111046137 A CN 111046137A CN 201911108881 A CN201911108881 A CN 201911108881A CN 111046137 A CN111046137 A CN 111046137A
- Authority
- CN
- China
- Prior art keywords
- characters
- emotion
- character
- facial
- emotional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/374—Thesaurus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Machine Translation (AREA)
Abstract
The invention relates to a multidimensional emotion tendentiousness analysis method, which comprises the following steps: (1) constructing a facial character dictionary; (2) extracting the character features and calculating a multi-dimensional emotional tendency value; (3) calculating a multi-dimensional emotional tendency value of the text and Emoji combined part; (4) calculating a multidimensional emotional tendency value of the network comment sentence: and (3) comprehensively considering the text, the Emoji combined part and the character part as for the emotional tendency of the network comment sentence, and obtaining the multi-emotional probability of the network comment sentence after weighting the emotional tendency values of the two parts obtained in the steps (3) and (4).
Description
Technical Field
The invention relates to a multi-dimensional emotion tendency analysis method for fusing emoticons and short texts on a network comment.
Background
The emotional behavior analysis of the social network plays an important role in a plurality of fields such as product comment, public opinion monitoring and information prediction. In the field of electronic commerce, emotion analysis can provide historical emotion evaluation abstracts of products for users, and the users can conveniently and quickly know product evaluation information. In the public opinion monitoring field, because people all have network speaking rights, various topics and viewpoints related to the national civilization can be published at any time, the direct influence of virtual social networks and real social interaction on the society is larger and larger, and the national security and the social stability are directly influenced. The method has the advantages that the emotional behaviors of people in the network are detected and analyzed, and the method plays an important role in maintaining national stability and promoting social development.
Some early methods only perform emotion analysis on short texts by using a rule-based and machine learning-based method, some works currently combine texts and Emoji to perform emotion classification, the characters are used as important emoticons in more and more media comments, and emotion analysis in a Chinese social network almost completely filters out the characters, so that emotional characteristics are lost, and the influence of the used characters on the social network comment emotion cannot be accurately judged.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, and provides a multi-dimensional emotion tendency analysis method for fusing emoticons and short texts aiming at social network comments, so that the individual personality structure of a user is enriched, the establishment of information such as user attributes, behaviors and emotion dynamic changes is facilitated, the emotion fine granularity analysis accuracy of a social network group is improved, and support is provided for public opinion analysis, social reasoning, rumor detection, privacy protection and the like. In order to achieve the purpose, the invention adopts the following technical scheme:
a multi-dimensional emotional orientation analysis method comprises the following steps:
(1) constructing a dictionary of characters
Constructing a facial character dictionary, wherein each emotion label comprises a series of facial characters, each facial character only belongs to one emotion label, determining body or facial expressions corresponding to the characters in the facial characters according to a human kinematics theory, defining the facial character emotion labels, neglecting the facial characters which cannot be deduced to have no obvious emotional tendency, and extracting the facial characters through a regular expression.
(2) Extracting character features and calculating multi-dimensional emotional tendency values
Extracting the length of the characters, the ASCII character ratio, the number of different types of characters and the number of characters with the largest occurrence frequency as structural characteristics through the constructed character dictionary; the sum of all character emotion category probabilities in the characters serves as a category characteristic and the emotion category probabilities of all elements in the characters serve as a kinematic characteristic, a characteristic vector is constructed by the extracted characters and serves as the input of a neural network to obtain a score vector of an emotion label, the multi-dimensional emotion classification probability of the characters is further obtained, and the probability serves as an emotion tendency value;
(3) calculating the multidimensional emotional tendency value of the combined part of the text and the Emoji
Splicing word vectors of the text and Emoji into a combined vector, respectively extracting local and global features through a deep learning model based on attention, improving the semantic expression capability of the part through feature fusion, and taking the fused feature as the input of another neural network to obtain a multi-dimensional emotion tendency value of the part;
(4) computing a multidimensional emotional tendency value of a network comment sentence
And (3) comprehensively considering the text, the Emoji combined part and the character part as for the emotional tendency of the network comment sentence, and obtaining the multi-emotional probability of the network comment sentence after weighting the emotional tendency values of the two parts obtained in the steps (3) and (4).
Detailed Description
The technical scheme of the invention is as follows:
(1) the method comprises the steps of constructing a Chinese character facial character dictionary on the basis of a Japanese character dictionary, determining a facial character emotion label in the construction process, determining body or facial expressions corresponding to characters in the character according to a human body kinematics theory, neglecting and deducing the character without obvious emotion tendency, extracting the character through a regular expression, and then cutting off an explanation language in the Japanese character.
(2) Calculating multi-dimensional emotional tendency of characters
Extracting the length of the Chinese characters, the ASCII character proportion, the number of different types of characters and the number of characters with the largest occurrence frequency as structural characteristics through the constructed Chinese character dictionary; the sum of all character emotion category probabilities in the characters serves as category characteristics and the emotion category probabilities of all elements in the characters serve as kinematic characteristics, extracted character characteristics are used for constructing characteristic vectors, and the characteristic vectors serve as input of a neural network to obtain score vectors of emotion labelsFurther obtaining the multi-dimensional emotion classification probability of the characters and the characters, and taking the probability as an emotion tendency value QK。
(3) Calculating the multidimensional emotional tendency value of the text and Emoji combined part
The text and the Emoji vectors are spliced into a combined vector, local and global features are extracted through a deep learning model based on attention, and the semantic expression capability of the part is improved through feature fusion. Finally, the fusion features are used as the input of the neural network to obtain the multidimensional emotional tendency value Q of the partTE。
(4) Computing a multidimensional emotional tendency value of a network comment sentence
For the emotional tendency of the network comment statement, a text, an Emoji combination part and a character and a word part are comprehensively considered, a positive variable parameter lambda is set, and the emotional tendency values of the two parts are weighted to obtain the multi-emotion probability of the network comment statement:
Q=λQTE+(1-λ)QK。
Claims (1)
1. a multi-dimensional emotional orientation analysis method comprises the following steps:
(1) constructing a dictionary of characters
Constructing a facial character dictionary, wherein each emotion label comprises a series of facial characters, each facial character only belongs to one emotion label, determining body or facial expressions corresponding to the characters in the facial characters according to a human kinematics theory, defining the facial character emotion labels, neglecting the facial characters which cannot be deduced to have no obvious emotional tendency, and extracting the facial characters through a regular expression.
(2) Extracting character features and calculating multi-dimensional emotional tendency values
Extracting the length of the characters, the ASCII character ratio, the number of different types of characters and the number of characters with the largest occurrence frequency as structural characteristics through the constructed character dictionary; the sum of all character emotion category probabilities in the characters serves as a category characteristic and the emotion category probabilities of all elements in the characters serve as a kinematic characteristic, a characteristic vector is constructed by the extracted characters and serves as the input of a neural network to obtain a score vector of an emotion label, the multi-dimensional emotion classification probability of the characters is further obtained, and the probability serves as an emotion tendency value;
(3) calculating the multidimensional emotional tendency value of the combined part of the text and the Emoji
Splicing word vectors of the text and Emoji into a combined vector, respectively extracting local and global features through a deep learning model based on attention, improving the semantic expression capability of the part through feature fusion, and taking the fused feature as the input of another neural network to obtain a multi-dimensional emotion tendency value of the part;
(4) computing a multidimensional emotional tendency value of a network comment sentence
And (3) comprehensively considering the text, the Emoji combined part and the character part as for the emotional tendency of the network comment sentence, and obtaining the multi-emotional probability of the network comment sentence after weighting the emotional tendency values of the two parts obtained in the steps (3) and (4).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911108881.8A CN111046137A (en) | 2019-11-13 | 2019-11-13 | Multidimensional emotion tendency analysis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911108881.8A CN111046137A (en) | 2019-11-13 | 2019-11-13 | Multidimensional emotion tendency analysis method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111046137A true CN111046137A (en) | 2020-04-21 |
Family
ID=70232672
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911108881.8A Pending CN111046137A (en) | 2019-11-13 | 2019-11-13 | Multidimensional emotion tendency analysis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111046137A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899298A (en) * | 2015-06-09 | 2015-09-09 | 华东师范大学 | Microblog sentiment analysis method based on large-scale corpus characteristic learning |
CN106446147A (en) * | 2016-09-20 | 2017-02-22 | 天津大学 | Emotion analysis method based on structuring features |
CN106776554A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | A kind of microblog emotional Forecasting Methodology based on the study of multi-modal hypergraph |
CN107515855A (en) * | 2017-08-18 | 2017-12-26 | 武汉红茶数据技术有限公司 | The microblog emotional analysis method and system of a kind of combination emoticon |
CN110020437A (en) * | 2019-04-11 | 2019-07-16 | 江南大学 | The sentiment analysis and method for visualizing that a kind of video and barrage combine |
-
2019
- 2019-11-13 CN CN201911108881.8A patent/CN111046137A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104899298A (en) * | 2015-06-09 | 2015-09-09 | 华东师范大学 | Microblog sentiment analysis method based on large-scale corpus characteristic learning |
CN106446147A (en) * | 2016-09-20 | 2017-02-22 | 天津大学 | Emotion analysis method based on structuring features |
CN106776554A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | A kind of microblog emotional Forecasting Methodology based on the study of multi-modal hypergraph |
CN107515855A (en) * | 2017-08-18 | 2017-12-26 | 武汉红茶数据技术有限公司 | The microblog emotional analysis method and system of a kind of combination emoticon |
CN110020437A (en) * | 2019-04-11 | 2019-07-16 | 江南大学 | The sentiment analysis and method for visualizing that a kind of video and barrage combine |
Non-Patent Citations (5)
Title |
---|
JIABO ZHANG.ET.L: "A Deep Learning Model Enhanced with Emojis for Sina-Microblog Sentiment Analysis", 《2019 IEEE INTERNATIONAL CONFERENCES ON UBIQUITOUS COMPUTING & COMMUNICATIONS (IUCC) AND DATA SCIENCE AND COMPUTATIONAL INTELLIGENCE (DSCI) AND SMART COMPUTING, NETWORKING AND SERVICES (SMARTCNS)》 * |
SHUO YU.ET.L: "Emoticon Analysis for Chinese Social Media and E-commerce: The AZEmo System", 《ACM TRANSACTIONS ON MANAGEMENT INFORMATION SYSTEMS》 * |
教育部语言文字信息管理司组编: "《中国语言生活状况报告 2014》", 30 September 2014 * |
谭皓等: "基于表情符注意力机制的微博情感分析模型", 《计算机应用研究》 * |
黎巎: "《旅游大数据研究》", 30 July 2018 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Islam et al. | Detecting depression using k-nearest neighbors (knn) classification technique | |
Deshwal et al. | Twitter sentiment analysis using various classification algorithms | |
Wu et al. | Chinese micro-blog sentiment analysis based on multiple sentiment dictionaries and semantic rule sets | |
CN111046136A (en) | Method for calculating multi-dimensional emotion intensity value by fusing emoticons and short text | |
CN105893582A (en) | Social network user emotion distinguishing method | |
CN107688576B (en) | Construction and tendency classification method of CNN-SVM model | |
Mehndiratta et al. | Identification of sarcasm using word embeddings and hyperparameters tuning | |
Asinthara et al. | Categorizing disaster tweets using learning based models for emergency crisis management | |
Sindhuja et al. | Twitter Sentiment Analysis using Enhanced TF-DIF Naive Bayes Classifier Approach | |
Du et al. | Sentiment analysis method based on piecewise convolutional neural network and generative adversarial network | |
Su et al. | An improved BERT method for the evolution of network public opinion of major infectious diseases: Case Study of COVID-19 | |
Yuan et al. | Sentiment analysis of microblog combining dictionary and rules | |
CN116663566A (en) | Aspect-level emotion analysis method and system based on commodity evaluation | |
CN111046137A (en) | Multidimensional emotion tendency analysis method | |
Islam et al. | Bengali social media post sentiment analysis using deep learning and bert model | |
Qamar et al. | Emotion detection from text using fuzzy logic | |
Li et al. | Automatic classification of interactive texts in online collaborative discussion based on multi-feature fusion | |
Ahire et al. | Fuzzy Approach for Context Identification into Ambient Computing | |
Chavan et al. | Machine Learning Applied in Emotion Classification: A Survey on Dataset, Techniques, and Trends for Text Based Documents | |
Sahu et al. | An Emotion based Sentiment Analysis on Twitter Dataset | |
Khan et al. | A Review on Sentiment Analysis of Twitter Data Using Machine Learning Techniques | |
Suraj et al. | Conversational assistant based on sentiment analysis | |
Sakhrani et al. | Contextualized Embedding based Approaches for Social Media-specific Sentiment Analysis | |
Zhang et al. | Object recognition base on deep belief network | |
Rai et al. | Tourism Review Sentiment Analysis using Lexicon Features and Machine Learning Approach |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200421 |