Emotion Analysis on Turkish Texts: a Systematic Mapping Study

Abstract

In recent years, the increase in internet usage also led to an increase in data. Therefore, the importance of classification has increased to understand and analyze the data more easily, and this made emotion analysis a necessity. Research in this area shows that emotion is a key tool in analyzing texts. However, there are very limited studies conducted on the text in Turkish language. The purpose of this study is to detect emotions in Turkish texts by analyzing the results of earlier studies in this scope. The following research questions are aimed to be answered: What are the years and count of publications What are the techniques used in the emotion analysis on texts What are the main data sets Which types of emotion are detected Which tools are used in emotion analysis For this reason, a systematic mapping study is performed to categorize and summarize the existing literature on emotion analysis. To ensure relevant papers in this field, an automated search process is performed for this study. Web of Science and Scopus databases are used for searching current and past research in emotion analysis on texts. There are 66 articles found on this subject. Then, by using filtering, irrelevant studies are removed. According to a set of inclusion and exclusion criteria, 9 relevant papers are selected in this field. The machine learning method is used for text analysis that automatically identifies the patterns and makes decisions about emotions. © 2022 IEEE.