Research On Recommendation Method Considering Commodity Repeated Purchase Cycle

With the continuous expansion of the scale of e-commerce,the number and variety of goods have grown rapidly,and there has been an increasingly serious phenomenon of information overload.Personalized recommendation systems have emerged as the times require.Although the current recommendation technology has made great progress,how to accurately grasp the user's purchase preference and improve the recommendation accuracy rate is still a problem that the recommendation system needs to solve urgently.In the recommendation practice,if you do not consider whether the user has purchased related types of merchandise and repeat buying cycles and other factors,it may cause some merchandise to be recommended at the wrong time.This not only occupies the recommended resources,but may also cause consumers trouble.Based on this,this thesis studies the improvement of the recommendation method from the perspective of repeated purchase cycles of goods.First of all,taking into account the law of the user's purchase of goods hidden in the repeated purchase cycle of the goods,a prediction method of the user's purchase behavior based on the repeated purchase cycle of the goods is proposed.The period of the repurchase period of the short repetitive purchase period purchased by the user is calculated by calculating the repeated purchase period of the item,so as to predict the category of the item that the user may purchase during this period of time.Secondly,the collaborative filtering of product recommendation methods is improved based on repeated product purchase cycles.A purchased product repurchase state variable is introduced on the collaborative filtering algorithm based on the user purchase record,and then the original collaborative filtering recommendation result is filtered based on the product repurchase state.The experimental results show that the prediction method of the user's purchase behavior based on the repeated purchase cycle of the product can effectively predict the category of the customer's purchase,and the accuracy of the prediction is high.In addition,the collaborative filtering product recommendation improvement method based on the repeated product purchase cycle can effectively predict the purchase behavior of the customer,and the improved collaborative filter product recommendation method is significantly improved in accuracy,recall rate,and comprehensive evaluation value.