Sentiment Classification of English and Hindi Music Lyrics Using Supervised Machine Learning Algorithms

Abstract

Finding music based on one’s mood is difficult unless it is manually classified and separated into distinct playlists. This is especially tough when the song is not in English due to varying lexical and syntactic styles. Our project employs textual sentiment analysis by testing various binary classifier algorithms - Random Forest, Naive Bayes, Support Vector Machine (SVM), and AdaBoost - to gauge which method is best for classifying English and Hindi language music lyrics into positive (happy) and negative (sad) sentiment.