Research And Application Of Customer Product Recommendation System In Mobile Mall

With the increasingly fierce competition in the mobile mall industry,how to realize personalized recommendation of products according to customer characteristics is an important content of accurate marketing in the current mall.At present,most shopping malls adopt the matching mode of customer product retrieval,product name and theme for product association recommendation,but this recommendation method does not consider the factors of customer product browsing time,and does not combine user interest with product type to realize product recommendation.This system analyzes the customer's interest points through the customer's basic characteristics,product orders,product browsing behavior and other factors,and realizes online product customer marketing by using recommendation algorithm,which improves the marketing accuracy.Based on the analysis of algorithms and recommendation systems at home and abroad,this paper completes the following research work:(1)The user characteristics are obtained through the user's participation in online product browsing behavior.There are two main ways to obtain: the first way is the user's registration information,that is,the personal information registered by the customer at the initial registration stage,including the customer's age,education background,work and other attributes;Another way to obtain customer characteristics is the information of products browsed or purchased by customers on the Internet.Through the information of products browsed or purchased by customers on the Internet,the related information of similar customers can be extracted and the interest points of customers' products can be obtained.So as to complete the research and analysis of the customer feature acquisition function,and realize the design of the customer influence relationship model based on the collaborative filtering recommendation algorithm of the user influence relationship.(2)In order to improve the accuracy of product recommendation,the system introduces the frequency weight coefficient based on the customer impact relationship model for different types of products,and forms a new customer impact relationship model.The interest characteristic value is calculated by coefficient,the customer product interest rating is realized,the accuracy of customer interest characteristic is improved,the customer demand is fully reflected,and the design of customer product interest model is completed.(3)On the basis of completing the customer product interest model,according to the size of the interest value,the interest degree between the customer and a single product is expanded to the interest degree between each product resource,the top N products are found out,the product recommendation of the customer Top-N is carried out,and the customer recommendation model design is completed.This paper focuses on the analysis of system requirements,so as to determine the use of recommendation algorithm to complete the establishment of customer recommendation model,and further design the system framework.Through the flow chart,timing chart and key algorithm,the system modules such as customer management,product retrieval and commodity recommendation are designed,and the system modules are realized through Java EE framework.Finally,by comparing and analyzing the different effects of collaborative filtering recommendation algorithm based on user influence relationship with other different recommendation algorithms,the system function and performance are tested.The test shows that the product recommendation system designed in this paper meets the needs of precise marketing of the company's products.