F1 스코어를 이용한 한국어 감정 지수 연구

Abstract

Digital companies are processing and managing their business relationships between their customers, suppliers and employees in digitalized way. These companies innovate the traditional social structure by applying digital technology to society. Enterprises build and utilize information and communication technology (ICT) as a platform, including cloud computing, artificial intelligence (AI), and big data solutions. Transformation into a digital company and using data analysis processes in an organization will create new values or services. In addition, analyzing the flow of data within an organization in an e-business environment is a necessary activity for corporate survival.<BR> The purpose of this study is to analyze the emotions of internal and external customers, which are diversified in atypical data analysis. Wheel of emotions(Plutchik, 1980) is used to classify emotions and prioritize emotion words within the categorized emotion sets. The Korean emotional word index process was studied by classifying and ranking the emotional words in the emotion set using the F1 score, a tool for measurement of effect. The emotion index was ranked in 200,000 comments using 190 generalized emotional words in the emotion set. Emotions such as anger and joy, fear and happiness can be identified through statistical values. In addition, it is confirmed that customer needs or product preference analysis can be applied to get more insights in the company"s point of view.