The Detecting Dynamics of Public Opinions in Social Media using thematic modeling methods

Abstract

Digital technologies have led to the formation of the new level of socio-cultural space, which is expressed in the permanent presence of the phenomenon of virtual reality in everyday life. It is the main motivating tool for social and political transformations of society, which consist in accelerating the logical bound “public mood à public opinion à social action of the masses”. Understanding these changes in social phenomena actualizes the problem of forming systems for detecting the current public mood, which can be the basis of feedback between the authorities and society, or monitoring public mood as a reaction to the social and political situation.

This study would work on topic modeling focused on the algorithm employing Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA). The data collection of news announcements, that were published between 2020 and 202, is used as the main data resours with unstructed text. The stages of preprocessing include cleansing, stemming, and stop words. The advantages of LSA are fast and easy to implement. LSA, on the other hand, doesn’t consider the relationship between documents in the corpus, while LDA does. This study shows that LDA gives a better result than LSA.