Collaborative filtering is a challenging and emerging field of research with a wide range of applications. Main applications especially being in recommender systems due to its simplicity and efficiency. There are several algorithms and techniques proposed in litera- ture for the same. In this project we implement a collaborative filtering algorithm which makes use of Gaussian Mixture models to cluster users and reduce sparsity in the original data matrix and build a new interaction matrix, followed by item to item Jaccard simi- larity based scoring to make predictions. We also evaluate our method on three popular public data bases (namely Movie Lens 100K, 1M and Netflix) and analyse the obtained results.

• Datasets used: Movie Lens 100K, Movie Lens 1M and Netflix 1200

• [Report]{https://rakeshpavan333.github.io/docs/sc_repo.pdf}

• [Base Paper] {https://ieeexplore.ieee.org/document/8808904}