Improving Data Sparsity in Recommender Systems Using Matrix Regeneration with Item Features

With the development of Web, users spend more time accessing information that they seek. As a result, recommendation systems have emerged to provide with preferred contents by filtering abundant information, along providing means exposing search results effectively. These operate based on user reactions items or various item features. It is known sparse datasets are less reliable because recommender according responses. Thus, we propose method improve dataset sparsity and increase accuracy prediction using features A content-based concept proposed extract category rates from user–item matrix preferences organize these into vectors. Thereafter, present filter extracted vectors regenerate input for collaborative (CF). We compare our approach conventional CF mean absolute error root square error. Moreover, calculate regenerated existing matrix, demonstrate dense than one. By computing Jaccard similarity between sets in matrices, verify distinctions. The methods confirm if used as input, denser higher predictive can be constructed when methods. validity was verified analyzing effect composed high average ratings performance. low comparisons Improvements approximately 16% K-nearest neighbor 15% singular value decomposition, three times improvement original matrices obtained. reconstruction performance recommendations.