Hot topic detection and evaluation of multi-relation effects

Abstract

With the growth of social media, Twitter has become one of the most popularly used microblogging communication platforms between people. Due to the wide preference of Twitter, popular issues in public, events like local or global news and daily life stories can immediately publish on Twitter. Thus, a substantial number of hot topics are created by Twitter users in real-time. These topics can exhibit every incident of everyday life. Therefore, detection of hot topics can be used in many applications such as observing public judgment, product recommendation, and incidence detection. In this paper, we propose a method for detecting Twitter hot topics and evaluate the effect of multi-relations such as retweets and hashtags on hot topics. The dataset was generated by fetching tweets for a certain time and location by using GetOldTweets3 API. Then using the LDA topic modeling algorithm the hot topics were identified for each multi relation. Finally, the effect of each relation is described by using the coherence scores)