Sentence Emotion Classification for Intelligent Robotics Based on Word Lexicon and Emoticon Emotions

Abstract

In recent years, the information of social network increases explosively, since more and more people like sharing their experience and giving the comments on some events to convey their attitudes by textual information which are rich in emotions. In this study, we examine the effectiveness of emoticon emotion distribution by sentence emotion classification based on word emotion lexicon and emoticon emotion distribution. We construct two word lexicons by using the word emotion intensities and frequency respectively, and calculate the emoticon emotion distribution by the co-occurrence of emoticon and sentence labels. Four experiments have been done, and result proves the emotion contribution of emoticon in emotion classification.