Opinion Mining and Classification of Music Lyrics Using Supervised Learning Algorithms

Abstract

Music Lyrics is an important and meaningful part of any song that are helpful in investigations and classification of opinion (sentiment) develop from it. Opinion mining is also referred as sentiment analysis is the field of data processing that is used to find out opinion of an author, user and subjectivity from text. In this work we are considering only the English lyrical part of a song. WorldNet knowledge is then incorporate to find out synonyms of words. The Goal of this research is doing a linguistic investigation of music lyrics whether these songs useful for listeners or not and classifying them with positive and negative fulfilled present in them. In Order to evaluate this words involve opinion(sentiment) have been investigate with using POS tagger and classifying them into mood categories using different machine learning algorithms(supervised) Random Forest, Gradient Boosting and Voting Classifier(including logistic regression, Decision Tree and SVM) and compare with different parameters.