Construction of E-commerce Users'Repeated Purchase Behavior Prediction Model Based on Collaborative Filtering Technology
Due to the strong randomness of users'browsing directions on e-commerce platforms and the huge amount of information they are interested in,it is difficult to predict users'repeated pur-chases.In order to enhance user experience and satisfaction,and improve the quality of e-commerce service,a prediction model of e-commerce users'repeated purchase behavior based on collaborative fil-tering technology was built.Obtain the historical data set of e-commerce users,extract the statistical,derivative and behavioral attenuation features that are strongly related to users'repurchase behavior from the user commodity and user commodity categories levels,and input them into the XGBoost algorithm to predict whether users have repeated purchase behavior.The collaborative filtering algorithm is used to establish a scoring matrix for users with repeat purchase behavior.The Pearson correlation coefficient is used to measure the similarity between the two users and determine the nearest neighbor set of users with repeat purchase behavior.According to the score of the nearest neighbor user on the target goods,the user's score on the target goods is calculated,and several goods with high scores are obtained to gen-erate a recommendation list,so as to realize the prediction of e-commerce users'repurchase behavior.The experimental results show that the model can achieve the prediction of users with repurchase behav-ior.When the number of CART trees is 500,the feature quantity is 60,and the length of the recom-mendation list is 3,the prediction performance of the prediction model is the most outstanding,and the user rating of the recommendation results can reach 9.8.