A Recommendation and Genre-Classification Model for Sound-Alikes

Sound-alikes of a particular musical piece are songs or musical pieces that sound similar to it. Apart from the legal nature of a topic like this, there is a need among music listeners to recommend songs that match the tone of the one currently being listened to, genre-wise and based on how similar it sounds to the currently-played music. This research explores the idea of recommending five sound-alikes of an input song based on cosine similarity and genre classification upon the GTZAN dataset, such that the user can then pick and select what music they would prefer to listen to next. A variety of machine learning models have been studied for genre classification upon the said dataset, and a conclusion has been reached regarding which model performs the best.