Evaluation of Movie Recommender Systems Using Hesitant Fuzzy Logic and Deep Convolutional Spiking Neural Networks

Within the entertainment industry, movie recommendation systems have become important tools that enable users to explore films in large catalogues in line with their tastes and experiences by suggesting content relating to their preferences. This paper examines the efficiency of an innovative approach that combines Hesitant Fuzzy Clustering (HFC) with Deep Convolutional Neural Networks (DCNNs) to improve accuracy of movie recommendations. The proposed system was validated on Rotten Tomatoes and MovieLens 10m dataset, which showed considerable enhancements compared to traditional approaches like collaborative filtering and matrix factorization for accuracy, precision and Mean Absolute error as well. An integrated model consisting of HFC and DCNN demonstrated an accuracy level of 98% for MovieLens 10m dataset while it was equal to 99% for Rotten Tomatoes. It further shows a decrement of Mean Absolute Error (MAE) when compared with previous Matrix Factorization along with neural network. It is therefore evident from these findings that the developed model is robust enough, thereby creating higher levels of user satisfaction through more accurate recommendations. Future research will extend the application of this technique to other areas and further refine it to be able to accommodate a wider range dynamic user preference.