EFFECTIVE MEASURES FOR THE CLASSIFICATION OF TWITTER TREND TOPICS

Abstract

Social media users offer ascent to social trends as they share about basic interests, which can be activated by various reasons. In this work, we investigate the sorts of triggers that start trends on Twitter, presenting a typology with following four composes: news, continuous occasions, images, and commemoratives. While past research has dissected trending themes in a long haul, we take a gander at the soonest tweets that deliver a pattern, with the point of ordering trends right off the bat. This would permit to
give a separated subset of trends to end users. We investigate and try different things with an arrangement of clear dialect autonomous highlights dependent on the social spread of trends to classify them into the presented typology. Our technique gives an effective method to precisely arrange trending themes without need of outside information, empowering news associations to find breaking news progressively, or to rapidly distinguish viral images that may improve advertising choices, among others. The investigation of social highlights additionally uncovers designs related with each sort of pattern, for example, tweets about
progressing occasions being shorter the same number of were likely sent from cell phones, or images having more retweets beginning from a couple of pioneers