Music Listener Mood Prediction from the Lyrics Using Machine Learning

Abstract

            Human Beings are vexed up with a variety of problems such as anxiety, stress, work tensions, and emotional outbreaks. This may be caused due to Job, family problems, responsibilities, and pressure from friends. During the battle with an emotional crisis, a human desperately searches for means to riddance the problem. Music, being one of the most popular means of entertainment, can help during such situations. It provides a way to express our feelings and to enhance our state of mind. The core part of music is the mood. Every situation we go through has a spirit associated with it. There are many songs that are written on emotions. Many public places, such as restaurants, tourist places, and cultural events, have a theme song in the background. This enhances the mood of the customers. In this regard, we are performing a mood classification of songs using lyrics alone. We are implementing Decision Tree and Random Forest models for the problem. The exploratory outcomes through training and testing the model show that music related to “happy” and “sad” states of mind can be anticipated with sensible accuracy dependent on features extracted from tune verses.