Filtering and Tracking of Spreading of Information on Social Networks Using a Combination of LDA, SVM, and Naive Bayes Models

Collecting and analyzing big amounts of data to extract useful data is a big challenge that we face in modern society. This presents great opportunities and challenges in the field of computer science research. If successfully analyzed and made sense of, data can help determine market trends, growth trends of an organization or stop the spread of information on social media. In this paper, the authors will conduct research on the theories of the Latent Dirichlet Allocation model (LDA), algorithmic Gibbs sampling, Support Vector Machine (SVM), Naive Bayes theorem, and the Waikato Environment for Knowledge Analysis (Weka). The authors also analyze and design the research system. This research will construct an empirical system to aid in the qualification and control of information on social media, detect implicit themes and potentially negative messages, trace the spreader of this news, and determine the speed of this news spreading. We aim to finalize a support system to aid with decision-making in the research that focuses on hot topics and development trends in the future and stop the spread of negative information and fix it.