Qualitatively Predicting User’s Music Preferences using Integrated Collaborative Filtering

Computer Engineering

Collaborative Filtering based recommenders are very popular in electronic commerce. Main challenges are the sparsity of the user-item rating matrix, scalability of the system and cold-start. In this paper, a new hybrid method has been proposed where user-based collaborative filtering and item-based collaborative filtering are combined together, which reduces sparsity problem. Clustering is applied as a preprocessing step to make the system more scalable. Item features based clusters and profile based clusters are introduced to generate the quality predictions. Experimental results show that combined user-based and item-based filters produce quality predictions while using clustering as preprocessing.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.