Feature and Sentiment Based Opinion Mining and Summarizing on Twitter

概要

論文の詳細を見る
Twitter is one of the most popular social network websites. Through the twitter platform, users share either information or opinions about personalities, politicians, products, companies, events. For a sudden event, there are thousands of tweets related popping out every second. This makes it difficult for a user to read them to manage public opinions and make an informed decision. In this research, we aim to mine and to summarize all tweets of a topic. Our task is performed in three steps: (1) mine topic features that users expressed their opinions on. (2) identify the sentimental words (a specific words or emoticon) in each tweet and decide whether each tweet (which may contains more than one sentimental word) is positive or negative. (3) produce a summary using the discovered information. This paper proposes several techniques to perform these tasks. Our experimental results with three different sets of tweets demonstrate the effectiveness of the techniques.
2013-02-21