A Kullback–Leibler divergence-based fuzzy C-means clustering for enhancing the potential of an movie recommendation system

Abstract

Recommender systems (RS) are systems that filter information and help users to choose products from a large amount of information available online. RS recommend satisfactory and useful products (items) like movies, music, books and jokes to target users that they are interested in. In collaborative filtering (CF) movie recommendation system, timeliness and accuracy are considered as an indispensable entity since it needs to aggregate the emotions, reviews and preference of users in an optimal manner for aiding them to determine suitable movies of their interest. The potential factor of accuracy and timeliness purely depends on the significance of the utilized CF methods with effective nearest neighbors for facilitating recommendation to the users. In this paper, Kullback–Leibler divergence-based fuzzy C-means clustering is proposed for enhancing the movie recommendation system. In this proposed KLD–FCM–MRS scheme, KL divergence-based cluster ensemble factor is included in the fuzzy C-means clustering methods for enhancing the stability and robustness in the clustering process. The improved sqrt-cosine similarity is also used to find the effective nearest neighbors for an active user. This proposed movie recommendation scheme is compared to the baseline approaches for investigation. The experimental results of the proposed technique confirmed a superior accuracy, recall value, mean absolute error with reduced time on par with the benchmarked schemes used for analysis.