Asset Publisher Asset Publisher

Return to Full Page

ali shariat bahadori

ali shariat bahadori


Ali Shariat Bahadori

master

 

Timeline Summarization in Twitter by Event Detection

In recent years, social networks had rapid growth and become a part of everyday life of most people. Among them, Twitter is the most popular micro-blogging service, which encourages people to share their feelings and thoughts. High volume of the content published every day in Twitter makes the analysis of real-world events hard, time-consuming and difficult for humans to follow. Therefore, researchers have focused on tweet summarization techniques to generate summaries consisting of minimum content with no redundancy while preserving generality. In this research, we present a system to analyze the most important events from users viewpoint and generate summary of main events to keep users updated about current events in a short time. Although, the focus of this research is Persian tweets posted by users, the proposed algorithms are language independent and they can be applied to other languages as well. The proposed approach consists of two main phases: The first phase models events in Twitter and detects main events, then tweets are clustered based on the event they represent. In the second phase, some existing tweet summarization methods are utilized to present a method for generating summary of important events. In this research we used unique features of Twitter including number of retweets, microblog structure and spam filtering to generate higher quality outputs in comparison to other researches. Keywords: Summarization, twitter social network, event detection