Research On Movie Recommendation System Based On Sparrow Search And Clustering Collaborative Filtering

As the world entering the era of big data,the data generated by the film industry shows an exponential growth.The growing number of users and films in the film platform makes it more and more difficult for users to find the films they are really interested in.In a variety of movie data,how to associate movies with users and help users obtain useful information is a big problem we are facing.With the deepening of the research,a personalized service system that can be customized according to the demand,the recommendation system,was born and applied to the film platform.Recommendation system can mine the needs of each user,find the types of movies that users are interested in,and customize personalized services for users.Collaborative filtering algorithm is favored by experts in various fields because of its strong adaptability.It is widely used in movie recommendation system.But it still has many shortcomings,such as the increasingly sparse scoring matrix with the increase of data,the lack of new data cold start and the ability of mining popular movies,which need further research.In view of the problems existing in the movie recommendation system,this paper does the following research.Firstly,K-means clustering algorithm and collaborative filtering algorithm are combined to solve the problem of sparse data.A new swarm intelligence optimization strategy,sparrow search algorithm,is introduced to optimize the clustering center point to solve the problems of K-means algorithm,such as sensitive initial center point and uncertain K value before operation.A video recommendation algorithm based on sparrow search and clustering collaborative filtering is proposed.The algorithm effectively alleviates the problem of falling into local optimum due to the selection of initial clustering centers is too concentrated.Then,aiming at the premature convergence of the sparrow search algorithm and the inability to dynamically adjust the step size,bird swarm optimization strategy and dynamic step weight strategy are used to improve the sparrow search algorithm.At the same time,after analyzing the limitation of the scoring matrix in the recommendation system,the user scoring preference model is used to optimize the scoring data in the original matrix.Finally,a movie recommendation algorithm based on scoring preference and improved sparrow search clustering is proposed.Furthermore,an experiment is designed on the real movie scoring data set,in order to compare with other related recommendation algorithms.The results show that the algorithm can improve the accuracy and efficiency of recommendation to varying degrees,and reduce the problem of data sparsity effectively.