Media system dependency theory: Emotional expression on Sina Weibo

Abstract

In order to guide the correct direction of public opinion on public events and create a harmonious and healthy social atmosphere, this paper uses media system dependency theory to study the current Sina Weibo platform, which has the highest national discussion, for emotional expression. In this paper, RNN recurrent neural networks and their improved algorithms GUR-LSTM and LSIN-LSTM are introduced for Sina Weibo user attention and self-attention mechanism and Sina Weibo user input layer construction. The algorithm model is then validated to ensure its validity, followed by full-text algorithm validation based on information from the database to clarify the formation of the main communication symbols and interaction ritual chains of Sina Weibo emotional expressions and the correlation between Sina Weibo emotional tendencies and emotions. The results show that "pictures" and "audio" are the least frequent communication symbols in the selection and use of Sina Weibo, accounting for 0.1% of the total, while "multiple symbols" is the most frequent symbol, accounting for 76.4% of the total. The top symbol, accounting for 76.4%, is "multiple symbols". At the significance level of 0.1, the asymptote of the emotional score model is 0.218, and the asymptote of the emotional media is 0.572, which means that the regression coefficients of each explanatory variable are significant. Therefore, the research results of this paper have credibility and can be applied to the study of Sina Weibo users' emotions, and then extended to other media field studies.