Assigning Polarity Scores to Facebook Myanmar Movie Comments

Abstract

User-generated texts such as reviews, discussions or comments are valuable indicators of users' preferences. Apart from binary classification (positive or negative) of the reviews, some researchers calculated polarity scores that give a very concise summary and provide more information of the reviews. In this paper, a system for assigning polarity scores to Facebook Myanmar movie comments is proposed. Myanmar is a language with underdeveloped electric resources. As this is pioneering work for this combination of language and sentiment analysis, the polarity scores of each positive and negative word in the movie domain-specific polarity lexicon is calculated. And then the polarity scores to each comment of the plain text movie corpus are assigned. The proposed system achieves 89% and 85% accuracy on positive and negative opinion words respectively in the evaluation of polarity score lexicon. We also make the comment polarity for 3-class evaluation and 5-class evaluation based on the scores of comments. General TermsSentiment analysis