The emotion recognition from Uyghur sentences based on combination of Class Discriminating Words and sentiment dictionary

Recently, emotion recognition has become a hot research topic in Natural Language Processing. Feature selection (FS) is a key process in Uyghur text emotion recognition, which will directly affect the accuracy of text emotion recognition. In this paper presents a recognition method for Uyghur sentence sentiments, such as angry, happy, sadness and wonder, based on combining Class-Discriminating Words (CDW) and sentiment dictionary. First, based on the characteristics of sentiment expression in Uyghur sentence text, features are extracted by using CDW feature selection method, and are used to emotion recognition. Second, some emotional words are collected by manually, and are built a sentiment dictionary, then combined with CDW feature words re-used to emotion recognition. The experimental results show that the method is effective in Uyghur text sentence emotion recognition. Therefore, it verifies the validity of this method.