K-Nearest Neighbor and Collaborative Filtering-Based Movie Recommendation System

Abstract

AbstractOver the past several years, the recommendation system has become one of the important aspects in our day to day life. Recommendation systems are used in various areas like YouTube, Amazon, etc. Collaborative filtering algorithms have been used in recommendation systems for a long time. They have been effective in resolving a number of issues with commercially available systems. Methods based on a user’s neighborhood have showed potential in forecasting user ratings. The aim of this paper is to design and evaluate ‘KNN algorithm and Collaborative Filtering algorithm’ for producing movie recommendations. The dataset used in this paper is ‘Movielens dataset’ which is downloaded from Kaggle. The system was implemented using ‘Python programming language’. Initially, compared different distance measures and then performed correlation. Based on the correlation value, developed system will recommend similar movies/users. Performances of both the developed algorithms were analyzed in terms of accuracy. Finally the result shows that the accuracy of the recommendations is very good and we will get more accurate movie recommendations based on the combination of “KNN algorithm and collaborative filtering algorithms”.KeywordsRecommendation systemKNNCorrelationCollaborative filteringCosine similarityMatrix factorizationMachine learning