The Research On Personalized Recommendation Algorithm Based On Bipartite Network Structure

In today's society,people receive massive amounts of information through various channels every day,which leads to information overload.The emergence of a personalized recommendation system has greatly eased this phenomenon.It can help people find what they need and recommend them to users from a vast amount of information.However,with the rapid development of the Internet,the information and resources in the network have grown exponentially,and the content structure has become more complicated.The traditional personalized recommendation system is difficult to meet the increasingly complex needs of people,making the personalized recommendation system face a huge The challenge.Therefore,the research and improvement of the personalized recommendation algorithm has great significance in the academic field and the commercial field.In many personalized recommendation algorithms,the recommendation algorithm based on bipartite graph network structure has received extensive attention from the academic community because of its high hit rate and no restriction on project types,and many scholars have improved and optimized this method.After an in-depth study of predecessors' work,an improved algorithm based on project attributes is first proposed.The algorithm uses the number of attributes of the project as an influencing factor to improve the first resource allocation method based on the bipartite graph network structure recommendation algorithm.On the one hand to ease the data sparse situation,on the one hand to improve the degree of personalization of the recommendation results.And simulation experiments for the user to recommend movies on the MovieLens dataset prove that the improved method has better performance in terms of diversity and enhances the personalized degree of the recommendation results.Although the improved algorithm proposed in this paper improves the degree of personalization of recommendation results,it does not improve the accuracy of recommendation results.Therefore,based on the project attributes,this paper proposes a combination of user preferences and project attributes on network structure recommendation algorithm.The algorithm uses the user's preference for project attributes and the similarity of user preferences to improve the entire resource allocation process,so that the distribution of resources between users and projects becomes more reasonable,resulting in better personalized recommendation results.Finally,a simulation experiment was also performed on the MovieLens data set to simulate the recommendation of the user to the movie,and it was proved that the improved algorithm performed well in terms of accuracy,diversity,and decrease in the popularity of the recommended item,and improved the accuracy and individualization of the recommendation result.And meet the individual needs of users.The main work of this article is as follows:(1)A bipartite graph network structure recommendation algorithm based on project attributes is proposed.(2)Propose a network structure recommendation algorithm that combines user preferences and project attributes.