The Research And Implementation Of Recommendation Algorithm Based On Bipartite Network

Faced with the rapid development of the Internet,the information data on the Internet has exploded,people need to have an effective way to get information.Moreover,personalized recommendation,as a means to help users to get personalized information,has been widely developed and applied.Among the numerous recommendation algorithms,bipartite network-based projection recommendation algorithm has received extensive attention because of its good recommendation effect,not limited by the type of items,and simple operation process.However,there are still some deficiencies in the existing recommendation algorithms based on the bipartite network,which need to be solved.This paper focuses on the bipartite network-based recommendation algorithm for the in-depth study and compares their advantages and disadvantages,and puts forward bipartite graph allocation model based on the user rating standard,and introduce temporal information to quantify the effects caused by the phenomenon of user interest shift,and also mix bipartite network-based recommendation model and network trust recommendation model together for multiphase algorithm to Ease the impact of data sparsity.Meanwhile the improved recommendation algorithm is applied to the music recommendation,in order to reflect the practicality of the algorithm.The main research results are as follows:Firstly,in this paper,we mix the bipartite network-based recommendation model and network trust recommendation model together.And to improve the existing bipartite network-based recommendation model,the algorithm model is integrated into the user's score standard information and time dynamic factors.Moreover the recommended results were validated and evaluated using the MovieLens data set.Through the comparison of RMSE and MAE of the algorithm,the improved algorithm has achieved good results.Secondly,except for improving the existing bipartite network-based recommendation algorithm,this paper designs and implements the Web application of music recommendation which apply the improved recommendation algorithm.The App includes the recently updated module,hit list module,search module and reviews praise module.