Tweet Summarization: A New Approch
Siddhi Naik, Shruti Lade, Swati Mamidipelli, Ashwini M. Save · 2018
Social Media has become a crucial source for gaining information about current goings-on in the world. Twitter is considered as fastest and most popular means of communication today where thousands of people tweet everyday regarding an event or a news-story. Number of tweets are posted on twitter on daily basis, which are analyzed by sentiment analysis to draw summary of opinions expressed by the user and categorize them to know the views of the user on an issue. The problem with sentiment analysis is that there are millions of users with differences in opinion to test. There are as well practical difficulties that arise while performing sentiment analysis. It may happen that someone may tweet some irrelevant stuff about an event and in such cases summarization comes into picture and plays a significant role. For providing user with better results tweets are not only summarized but they are first segmented. Segmentation helps in conserving semantic meaning of tweet. For summarization of tweets there are various clustering algorithms available which different researchers have used in their systems like K-means, Ant Colony. One of the optimal algorithms is the Particle Swarm Optimization Algorithm (PSO). PSO works in an analogous way as that of swarm of bees. And hence to provide user with better results the proposed system makes use of the PSO algorithm for clustering of tweets for summarization.