Collaborative Filtering-Based Film Recommendation Technique Utilizing Time and Film Genres
Huilin Wang, Jianhua Wu · Journal of Physics Conference Series · 2020
Abstract Due to the uncertainty of users’ preferences for different movies at different times, movie recommendation poses a huge challenge. We present an improved collaborative filtering algorithm for film recommendation whose design was motivated by a need to extract meaningful information for further recommending from film viewing applications. To this end, the method we propose utilises an improved method for analyzing user interest in film genres, which can then be subsequently used to predict users’ ratings of unwatched movies in combination with the current viewing time. In our new algorithm, we extract genre information considering time decay factor and combine it with matrix factorization collaborative filtering algorithm, and propose a novel algorithm structure, the MFTGICF algorithm. Compared with general traditional collaborative filtering algorithms and some other methods, this algorithm has more precision and higher stability. Both the theoretical analysis and extensive experiments show the improvement of the effectiveness and efficiency of the proposed method.