한국어 감정 디지털 온톨로지 구축에 관한 연구

Abstract

The purpose of this study is to establish basic emotions and to understand the categories and characteristics of emotional words in order to construction a Korean emotion ontology. As artificial intelligence technology has developed, many studies have been conducted to understand human emotions using artificial intelligence technology. It was used to understand the consumer’s intentions through exploring emotions in movie reviews and product reviews. However, most of prior research has focused on two types of emotions (i.e., positive and negative emotions) and the roles of discrete emotions have yet been explored. Constructing an emotional ontology is one of the first steps to investigate discrete emotions using artificial intelligence technology. In this study, data for developing a discrete emotion ontology were collected. First, 24 basic emotions were set, and then emotions were tagged for all words in the Korean dictionary “Urimalsaem”. The set of this data is meaningful as basic data for Korean emotion recognition. In addition, it can be used as basic data for consideration of the emotional vocabulary of Korean.