Research On E-commerce Personalized Recommendation Method Based On User Trust Mechanism

With the era of big data coming,personalized recommendation has become an important research field of the electronic commerce website,which can quickly find and match user's preference information.User trust has an important effect in the recommendation system.However,the traditional E-commerce websites rarely consider the impact of user trust.In this paper,the user trust mechanism is integrated into the electronic commerce recommendation method.This paper mainly does the following work:1.To address problems such as data sparsity and low accuracy commonly found in the recommendation system,we add user trust to the recommendation algorithm.In view of the common score items in User-Item Rating Matrix,the total score and the proportion of the total score are considered,and asymmetric trust matrix is established.For the non common score project,we use the attribute information of items and the scoring weight to calculate the user's preference.Then,a new recommendation algorithm based on user trust is proposed.Experiments are carried out on the data set with many different kinds of recommendation algorithms.The results show that the proposed method has good performance in the accuracy,the recall rate,the mean absolute error and the root mean square error.2.In order to predict ratings accurately,the user trust and matrix factorization is combined,user score trust and user preference trust are integrated into the neighborhood model.At last,the results are compared with a variety of different recommendation algorithms.The results illustrate the proposed method is much higher in accuracy.3.The algorithm of graph clustering is improved,we consider the method of select the initial center point and the orphaned user,then we combined it with user trust to apply to the recommendation.Finally we compared our method with other related recommendation algorithms.The results show that the proposed method has good effects in both accuracy and recall,and it can be used in the electronic commerce recommendation system.