Construction Of User Clustering Movie Recommendation System Based On Bipartite Graph Networks

The rapid development of information technology has brought people a lot of convenience,with the increase of Internet users,the user can't select the appropriate data in generating massive amounts of data,and there can't be a lot of commodity information lets the user pick.The birth of the recommendation systems to ease the information overload to the user's problems,and the data sparsity and cold start of the system is also placed in front of the system.The sharp increase of the product and the user and the user's evaluation of the product is limited to the problem of data sparsity.In this paper,based on the comprehensive analysis of the current research status,the data sparsity is further studied.Firstly,a user clustering recommendation algorithm based on bipartite graph networks is proposed,and a number of user sub networks are integrated by clustering technology,in the user sub network to build a bipartite graph network when the use of material diffusion method to establish the weight matrix of the items,the weight matrix of the weighted sum of the user to calculate the value of the items recommendations.Finally,we analyze the correlation between the sub networks in the constructed sub network and record the relationship by using the matrix,and construct a user clustering recommendation algorithm based on the bipartite graph networks.Secondly,a recommendation algorithm based on user clustering and item classification is proposed.Firstly,the algorithm classifies the film according to the type of film,and then whether the user has seen this type of film to calculate the degree of interest in the genre of the film,the similarity between users in each sub network is calculated by combining item classification and user interest degree in clustering,and then all the similarity is weighted by the user's interest to get the final similarity.Finally,the correlation between the sub networks is analyzed,and the recommendation algorithm based on user clustering and item classification is constructed.Finally,the above two algorithms are tested on the data sets with different degrees of sparsity and compared with other existing algorithms.Experimental results show that the two algorithms are better than the contrast algorithm,and on the sparse degree of different data sets,the algorithm of this paper has little effect on the quality of recommendation.