A Technique to Detect Music Emotions Based on Machine Learning Classifiers

Abstract

Music has the power to evoke emotional responses and is a vital component of human life. Emotion Recognition of Music is useful for music comprehension, retrieval, and other music-related tasks. It is also important to be able to detect a person's emotional state through their voice. In this research, we suggested a technique for recognising song emotion that might also be used to recognise speech emotion. Various musical features are retrieved and throughout the process, data is fed into machine learning classification algorithms: Random Forest, SVM, Decision Tree, and Naive Bayes. When compared to other algorithms, the audio is analysed for emotional content and identifies six emotions (angry, calm, fearful, happy, neutral, and sad), with Random Forest having the best accuracy and performance. By increasing the number of features extracted and reducing noise, this method can be utilised to detect speech emotion in the future.