Emotion dataset from Indonesian public opinion

Abstract

An opinion is a type of judgment or a person's point of view about something. Twitter is a popular social media platform that includes a lot of public opinions and would be a suitable location to mine data in text form. With its vast population and active Twitter user base, Indonesia has the potential to be a source of opinion data mining. An opinion may be processed and result in the form of a person's emotional response towards something, such as whether they like, hate, love, or are happy about it. Upon that basis, a dataset of Indonesian-language tweets conveying public opinion on various topics was formed. The fact that there are only limited publicly available emotions text datasets in the Indonesian language supports our basis in this research to form our emotion dataset. The gathered data was cleaned and normalized in the pre-processing stage to the necessary form for study on the task of classifying emotions in Indonesian. The data collected is annotated with six emotional labels: anger, fear, joy, love, sad, and neutral.

Keywords

• Emotion classification
• Dataset
• Tweet
• Indonesia