Sentiment analysis of informal Malay tweets with deep learning
Ong Jun Ying, Muhammad Mun’im Ahmad Zabidi, Norhafizah Ramli, Usman Ullah Sheikh
Abstract
<table width="593" border="1" cellspacing="0" cellpadding="0"><tbody><tr><td valign="top" width="387"><p>Twitter is an online microblogging and social-networking platform which allows users to write short messages called tweets. It has over 330 million registered users generating nearly 250 million tweets per day. As Malay is the national language in Malaysia, there is a significant number of users tweeting in Malay. Tweets have a maximum length of 140 characters which forces users to stay focused on the message they wish to disseminate. This characteristic makes tweets an interesting subject for sentiment analysis. Sentiment analysis is a natural language processing (NLP) task of classifying whether a tweet has a positive or negative sentiment. Tweets in Malay are chosen in this study as limited research has been done on this language. In this work, sentiment analysis applied to Malay tweets using the deep learning model. We achieved 77.59% accuracy which exceeds similar work done on Bahasa Indonesia.</p></td></tr></tbody></table>