基于词典的中文微博情绪识别

Emotion Analysis of Chinese Microblogs Using Lexicon-based Approach

微博等社交媒体已成为表达个人情绪和感受的重要平台。自动分析微博文本表达的情绪对于迅速了解大众情绪走向以及调节个人情绪有着重要的意义。文中首次针对中文微博中的情绪进行自动分析,识别微博表达的喜、哀、怒、惧情绪。提出以词典为依据的基于规则的方法,通过实验详细分析了中文情绪词典在社交媒体文本分析中的现状,讨论了存在的主要问题。并深入讨论了微博中情绪表达的语言特点,为建立高精度的情绪分析系统提供了依据。

The proliferation of microblogs has created a digital platform where people are able to express themselves through a variety of means. Automatic analysis of the emotional content in microblogs plays an important role in capturing popular feelings and adjusting personal mood. In this paper, a lexicon-based approach was proposed to automatically determine whether a microblog expresses one of the four basic emotions：joy, sadness, anger,and fear. We performed an extensive analysis of current Chinese emotion lexicons to understand their roles in analyzing social media text. The experimental results show that lexicon is a crucial resource in emotion analysis. The results also reveal limitations of current Chinese emotion lexicon. The characteristics of emotion in microblgs are identified for building advanced emotion analysis system.