A New Variant of Clustering based Normalized Knn Movie Recommendation System

In recent years, Recommender Systems (RS) have experienced significant growth across various sectors. Because of the recent growth of many e-commerce companies and online video services like Netflix, YouTube, Hotstar, etc., researchers are quite interested in it. The goal of the collaborative filtering-based recommender system is to provide users with movie or video recommendations based on their past viewing preferences. A rating matrix is typically used to display this data. In this study, the K-nearest neighbor and hierarchical clustering techniques are used to create a movie recommendation system. This work develops recommender systems using a range of instruments and methods. Several techniques, such as clustering techniques, KNN, collaborative filtering, and content-based filtering, are described in depth. The suggested approach for developing the movie recommender system is then illustrated, showing how it would be implemented and operated. First, the proposed model checks similar movies in the movie list and uses a hierarchical clustering technique to make the cluster of similar movies. For this process, Jaccard distance is used. Then the proposed methodology uses the K-nearest neighbor algorithm to recommend the best suitable movies to the users. To verify the suggested method, this study uses the Movielens dataset, which is available on Kaggle. Python is used as the programming language for implementing the system.