Big Data Analysis of Taiwan-related Public Opinion in Youth Online Communities of Chinese Mainland:Based on LDA Model and Text Sentiment Calculation
This paper combines big data methods such as LDA modeling and text sentiment calculation with traditional data processing methods to analyze the public opinion texts in mainstream youth online communities in the mainland.It finds that the current expressions of Taiwan-related issues among these youth online communities can be categorized into five major topics:"politics""entertainment""confrontation""life"and"communication".In terms of content characteristics,the topics of"politics"and"communication"are more likely to gain viewer support;the"entertainment"topic shows a"fan-oriented"characteristic;the"confrontation"topic is more likely to receive short-term attention and retweets;while the"life"topic is more likely to generate extensive discussions.In terms of sentiment characteristics,Taiwan-related discussions are characterised by a"love-hate"dichotomy,with"praise"and"criticism"being the most frequently expressed sentiments.However,among the positive expressions,about 20%-25%are sarcastic or ironic.In terms of interaction forms,negative and intense expressions are more likely to generate attention and discussion.In addition,the gender of users is significantly correlated with the sentimentality of Taiwan-related discussions.In the future,it is necessary to further strengthen the tracking of Taiwan-related public opinion,pay attention to the guidance of public opinion,and consolidate the construction of mechanisms for Taiwan-related communication and positive emotional expressions within youth online communities.
youth online communitiesTaiwan-related public opinionLDA modelsentiment calculation
NETL
NSTL
万方数据
