Movie Recommendation Algorithm Based on Ensemble Learning

With the rapid development of personalized services, major websites have launched a recommendation module in recent years. This will recommend information you are interested based on your viewing history and other information, thereby improving economic benefits website increasing number users. paper has introduced content-based algorithm, K-Nearest Neighbor (KNN)-based collaborative filtering (CF) algorithm singular value decomposition-based (SVD) algorithm. However, mentioned algorithms all for certain aspect, do not realize specific movies input by users which cause recommended content to deviate from need users, affect experience using. Aiming at this problem, combines above proposes three ensemble algorithms, KNN + text, user movie KNN, decomposition. Compared with traditional matrix factorization, method we proposed can make more recommendations deal problem cold start sparse processing issues extent.