The Sparse Matrix and Collaborative filtering system.

Abstract

In this thesis, we addressed to the sparse matrix problem in collaborative filtering. This problem happens when the matrix contains a few known data and there are many algorithms try to solve this problem, we choose two Different algorithms witch are the state-of-the-art in the collaborative filtering witch are: KNN and MF for predicting the messing values in the User-item Matrix to give recommendation to the end-user and according to the RMSE value we see that the MF with SGD is the best algorithm to predict missing value. In our next work we will try to develop an algorithm where the RMSE value is less than 1 when we applicate it to the dataset.