An Exploration of Semantic Features for Categorizing Emotion Verbal Corpus in Thai

Abstract

The rapid development of Artificial Intelligence has put forward higher requirements of Thai language for technological and innovative development in Thailand. As natural language processing needs the foundation of Thai language understanding, particularly the semantic analysis, the main purpose of this paper is to explore semantic features of Thai emotion words to find the standard criteria of emotion categorization in Thai language. This research paper takes the data on emotion wordlists from Thai dictionaries and thesaurus in addition of Thai National Corpus (TNC). Linguistic framework of Natural Semantic Metalanguage (NSM) is employed as the guidance of analysis and describing semantic attributes to build the semantic prime set of basic emotions in Thai. The result of this paper highlights the semantic features of the word ‘aai’ (shame or embarrassment) which reflect various classes of emotion. The example of analysis emphasizes that linguistic perspectives, semantic primes of NSM in particular, can be an appropriate tool for distinguishing and categorizing emotion terms and, in turn, for Thai emotion corpus creation.