Opinion Mining and Sentiment Analysis in Social Networks: A Retweeting Structure-Aware Approach

Abstract

Microblogs have become quick and easy online information sharing platforms with the explosive growth of online social media. Weibo, a Twitter-like microblog service in China, is characterized by timeliness and interactivity. A Weibo message carries the user's views and sentiments, particularly forms a fission-like spreading structure while being retweeted. Such structure accelerates information diffusion, and reflects different topics and opinions as well. However, current researches mainly focus on sentiment classification, which neither efficiently combine tree-like retweeting structure nor analyze opinion evolutions with a holistic view. In light of this, we build an opinion descriptive model, and propose an opinion mining method based on this model. With a microblog-oriented sentiment lexicon being constructed, a lexicon-based sentiment orientation analysis algorithm is designed to classify sentiments. Finally, we design and implement a prototype which can mine opinions with respect to retweeting tree structures and retweeting comments.