Affective Computing Based Chinese Emotion Trigger

Affective computing in artificial intelligence covers three kinds of emotion skills - emotion perception, emotion judgment, the ability for expressing emotions. Emotions could be identified by facial movement, voice and language. At present, most researches of emotion judgment are based on the first two methods, since feelings are usually expressed through voice, face expression, body movement and even language, which contain large amount of emotional information. However, emotions can also be conjectured by text when voice and expression pictures are not available. With the repaid growth of personal computers, interactive systems such as chat robot, games also become popular. Therefore research of emotion judgment by text turns hot. This thesis focuses on the research of text-based emotion.This study focuses on the way people judge emotion, and simulates this process on computer. The emotion trigger is contributed for the purpose of offering computer the foundation of emotion judgment. The emotion trigger refers the movements and event list. This paper presents an emotion judgment approach based on Chinese Emotion Trigger Dictionary. Our corpus consists of 500 Chinese sentences extracted from Email in ptopa Inc. The result shows that this method is reasonable and effective.Innovation points:1. This research of emotion judgment focuses on text, while most researches currently are based on face expression and voice.2. Emotion is judged by emotion trigger. For example, some could be "happy" if an emotion trigger "get a gift" is found in text.3. Not only emotion but also emotion intensity can be identified.