Social Emotion Analysis System for Online News

Peng Nie, Xue Zhao, Li Yu, Chao Wang, Ying Zhang

Abstract: Social emotion analysis of online users has become an important task for mining public opinions, which aims at detecting the readers' emotions evoked by online news articles. In this paper, we focus on building a social emotion analysis system (SEAS) for online news. The system has implemented a text data crawler for mainstream online news websites, the modules of document preprocessing, document representation, and also integrates successful emotion analysis methods and provides the corresponding performance evaluation. SEAS will automatically analyze the emotions towards certain news articles and output the predicted emotions and probabilities of being classified into these emotion categories. The experiments on the real dataset from online news service demonstrate the high practicability and reliability of SEAS.