Classification of Music Genre Using Machine Learning

Abstract

Deep learning approach is the first approach for classification of music genre using machine learning and for the purpose of predicting the genre label of a signal,a CNN model would be trained usually from end-to-end, spectrogram is used to carry out this process. Hand- crafted features were used in the second approach which is also the last approach. Four out of many traditional machine learning classifiers will be trained beside these features and their performance would be compared afterwards. Identification of features that would help the classification of this task needs to be performed. The databases of online music are growing day by day and nowadays it is very simple to access online music, due to this reason people are finding extremely uneasy to regulate the songs that they like to listen to. Nowadays music has become a very salient proportion of the Internet content, the most salient source of pieces of music is the net.