The Implantation Of A Commodity Recommender System In A E-commerce Platform

As the scale of the company continues to expand,the existing model of offline sales can not meet the demand.In order to keep up with the current trend of Internet technology,the company needs to develop an e-commerce platform suitable for its own.Users in this e-commerce platform,facing the dazzling products,hope that the platform can provide users with the needed goods quickly and pertinence.The emergence of recommendation system is to provide different display of goods to meet different needs of users.This article is intended to build a recommendation system that can improve the user experience.This recommendation system is on the original e-commerce platform and provides personalized recommendation service for users.Firstly,the thesis introduces the recommendation system,explains the different recommendation systems,compares their respective advantages and disadvantages,and explains the evaluation method of the recommendation system.Finally,it selects the recommended algorithm,points out the algorithm selected in this thesis,and expounds its rationality.Secondly,the requirements of the recommended system for the platform are analyzed,and the database of the recommendation system is designed and applied on the basis of module design and function design.Then the algorithm is designed according to the function of the recommendation system in different pages.It mainly defines how to classify the goods and how to calculate the user's interest in the different goods in the platform after obtaining the user's behavior characteristics.For different application scenarios of different pages,this thesis designs different recommendation algorithms,such as "seen and seen",which is based on the recommendation of a group user's commercial browsing behavior,which can be placed on the product details page(according to the details page),or the user center page(according to the browsing record)model.In the block,the algorithm is designed for different users' needs,so as to achieve different recommendation purposes.Finally,through the analysis of the system testing and sales data,the system availability,user interface,platform operability and other aspects are verified,and the feasibility of the recommendation system is proved.