Emotion detection on Myanmar texts

Abstract

At this age, World Wide Web is growing faster. Many companies have built and launch social media networks. People so widely use social media to get the latest news, to express their emotions or moods, to communicate with their friends and so on. Emotions of social media users are needed to analyze in order to apply in many areas. Many researchers do research on emotion detection using different techniques with their languages. Currently, there are no emotion detection systems for Myanmar (Burmese) language. So, this paper describes the emotion detection system for Myanmar language. This system uses our pre-constructed M-Lexicon, a Myanmar word-emotion lexicon, in the detection process. This system detects six basic emotions such as happiness, sadness, anger, fear, surprise, and disgust. In order to determine certain emotion from the text, we also apply rule-based decision making on sentence nature. We use Facebook users’ status, which has been written in Myanmar words. Emotions of user groups are also summarized in this system. Our approach achieves 86% accuracy for emotion detection in Myanmar texts.