We used Million Song Dataset provided by Kaggle to find correlations between users and songs and to learn from the previous listening history of users to provide recommendations for songs which users would prefer to listen most in future. Due to memory and processing power limitations, we could only experiment with a fraction of whole available dataset. We have implemented various algorithms such as popularity based model, memory based collaborative filtering, SVD (Singular Value decomposition) based on latent factors and content based model using k-NN. Memory based collaborative filtering algorithm gave maximum mean average precision. We believe that content-based model would have worked better if we would have enough memory and computational power to use the whole available metadata and dataset.
We have to learn some diffrent libraries also about dbms. and the power of Ai
Technologies used
Discussion