A Novel Hybrid Music Recommendation System using K-Means Clustering and PLSA

We propose a hybrid approach of music recommendation according to which user will get recommendations of the songs searched by the other users alike to the existing user on content basis and also songs listened by that similar set users on context basis. Here we will try to find the similar users on the basis of context using K-Means clustering technique and among those users we will find the most frequent or most listened songs on the basis of content using PLSA technique. To fulfil this task we have created a dummy dataset and taking assumption that it is valid for experimental point of view. In order to evaluate the utility of recommendations produced by using proposed approach of music recommender system, we have computed three metrics i.e. Precision, Recall and F-1 Score. The result of proposed methodology shows the promising result as evaluated by parameters i.e. precision, recall and F1-score on dummy dataset. It can be tested for the real dataset of users from future perspective which will definitely going to take some time because of availability of such data.