Complex emotion categorization and tagging for Chinese

摘要

To understand emotion and make machine emotion is one of the goals of affective computing. In order to understand language from interface of machine, both the meaning and the emotion are necessary to be interpreted correctly. But until now the study of fine-grained theory of emotion is still full of challenges. In this paper, a novel method to analyze emotion category is proposed according to the statistics of affective word hierarchy in Dictionary of contemporary Chinese. The emotion hierarchy is called complex emotion. Firstly, over 7,000 common affective words have be annotated and their detailed explanations had been collected for an affective lexicon, then we analyze the consistent relationship in the affective lexicon and consequently 52 salient complex emotion state hierarchies are categorized and tagged by a lexical affective clustering algorithm. The complex emotions are compared to the traditional definitions of basic emotions in psychology and have been evaluated to be valid by kappa value in the experiment. Moreover we also have tagged the semantic orientation for the collected words.