Font Size: a A A

Implicit Sentiment Analysis Of Users Based On Large-Scale Political Media Data

Posted on:2019-08-31Degree:MasterType:Thesis
Country:ChinaCandidate:Y X LiuFull Text:PDF
GTID:2416330626452404Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Government's credibility has gradually become a research hotspot,the Internet has developed rapidly in the recent years,and the number of Internet users has increased dramatically.The study of credibility of the government in the network society cannot be ignored.It is often that some core figures dominate the sentiment trend of the whole group in the network.The sentiment analysis of the core characters can reveal the sentiment of some groups about the government in the network society,and facilita te the government to take more targeted measures to enhance its credibility.This paper uses sentiment analysis techniques and key point recognition techniques of complex network,and combining with semantic representation model to conduct sentime nt analysis on key users and mine users' implicit sentiment.Specific work and innovat io ns are as follows:Firstly,the "structure + context" sentiment calculation model is proposed,and combining with the complex network,this model is used to conduct an empirica l analysis of the Tianya Forum data.The network about the relationship between users is established,and the key point recognition and the sentiment analysis about some key users based on the “structure + context” sentiment calculation model are conducted,which can reveal users' sentiment tendencies towards events related to the government from individual angle;some recommendations are given to the government to help it to take measures to enhance its credibility.Secondly,the “implicit sentiment analysis” model based on semantic representation is proposed,and an empirical analysis of Wikipedia data is conducted with the help of this model.The Chinese word vectors are trained,and these vectors are partitioned and calculated in vector space.The effect of semantic representation model on Chinese corpus is explained,and the implicit sentiment analysis of Wikipedia users from the whole angle on Chinese and American is realized.This paper mainly analyzes the implicit sentiment of social media users from two angles: the individual and the whole.The empirical analysis of Tianya community data shows that users generally pay more attention to economic and legal issues,and have more negative sentiments.The empirical analysis of Wikipedia data shows that the semantic representation model and IAT can be applied to the Chinese corpus to achieve better results,and there are some obvious differences between the views of the users to the Chinese and American on some terms.The above empirical analysis provides theoretical support for the government to enhance its credibility of the network society.
Keywords/Search Tags:Government's credibility, Key point recognition, Sentiment analysis, Implicit sentiment
PDF Full Text Request
Related items