Collaborative Filtering Recommendation Method Based On User Influence And Potential Factors

In recent years,with the rapid development of Internet,network information is expanding.Each of these data is growing rapidly and accumulating on a large scale.The growth of these resources is far more than the number of users.So in the face of massive data,it becomes very difficult for users to find information which they are interested in.In order to be able to lock the information quickly and accurately,the recommendation system emerges as the times require.Collaborative filtering is the most widely used technology in recommendation technology.Analyze the known preferences of a group of users to predict the unknown preferences of other users,this is the core of collaborative filtering.In the research of collaborative filtering,Due to the situation that people only consider explicit data or implicit data,the problem of data sparsity is caused.Although there are several widely used methods of similarity calculation,there is no consideration of the influence of the wind users and the change of time.The effect of the recommendation is not ideal,and the extensibility is poor.Therefore,this paper proposes a collaborative filtering recommendation method(UIPF-CF)based on the user influence and potential factors.Compared to other collaborative filtering recommendation methods,the method considered the explicit data and implicit data.And it introduced contribution factor and the time attenuation function.Break through the traditional method based on the graph model and alleviates the sparsity of the scoring project.It has a prominent contribution in improving the coverage,accuracy and extensibility of the recommendation.Finally,we verify the method proposed in this paper and the traditional collaborative filtering recommendation methods in the real dataset.The results show that the method proposed in this paper is superior to other methods on some indicators.The experiment proved the effectiveness of the method.