Enhanced Sentiment Classification for Informal Myanmar Text of Restaurant Reviews

Abstract

Nowadays, users' desire reviews and online blogs sites to purchase the products. With the rapid grown in social networks, the online services are gradually more being used by online society to share their sight, opinion, feelings and incident about a particular product or event. Therefore, customer reviews are considered as a significant resource of information in Sentiment Analysis (SA) applications for decision making of economic. Sentiment analysis is a language processing task which is used to detect opinion articulated in online reviews to classify it into different polarity. Most of resources for sentiment analysis are built for English than other language. To overcome this problem, we propose the sentiment analysis for Myanmar language by considering intensifier and objective words to enhance sentiment classification for food and restaurant domain. This paper aims to overcome the language specific problem and to enhance the sentiment classification for informal text. We address lexicon-based sentiment analysis to enhance the sentiment analysis for Myanmar text reviews and show that the enhancement of sentiment classification improves the prediction accuracy.