Intelligent Clothing Recommendation System Based On Hybrid Recall Model

With the development of artificial intelligence,big data,cloud computing and other technologies,e-commerce service websites have gradually stepped into a mature stage.Timely,accurate and customized recommendation services can not only predict users' consumption needs and solve the problem of information overload,but also maintain users' loyalty and improve the stickiness of the website platform.Since 2012,apparel products have become the number one selling product for online shopping,and the turnover has maintained steady and rapid growth for seven consecutive years.However,due to the special attributes of gender,style,popularity,etc.,personalized recommendations suitable for books,music,and information cannot be directly used in clothing products,but nearly half of women will pay attention to personalized recommendations on shopping platforms.Therefore,the research on personalized recommendation of clothing has very important practical significance.According to the component modules and basic framework of e-commerce personalized recommendation system,this research focuses on the section of recommendation method,and explores the key issues such as the establishment of user model.Combined with the marketing characteristics of apparel products,this topic has planned and discussed the user information collection,similar crowd delineation,the perspective dimension of behavioral preferences,and initially constructed a mixed-mode apparel intelligent recommendation system technical framework.This model establishes different types of recommendation models based on three recommendation techniques.Comprehensive consideration of factors such as demographic information,similarity of user preferences,differences in project preferences,and utilization of user historical behavior data.Effectively avoids the shortcomings of single recommendation technology and improves the accuracy of e-commerce system prediction of recommended clothing products.In order to verify the effect of recommendation,the mixed-mode was applied to J company's e-commerce platform fashion goods-men's wear online recommendation activity,demographic information was used to construct user profiles and 4A behavioral preference was used to establish user labels,finally implement recommendations.According to the user feedback information and related data indicators of the e-commerce background,the number of people in this activity has increased significantly.Therefore,the intelligent recommendation system based on mixed-mode has been proved to be able to effectively implement the recommendation of apparel related products,and can be widely used in e-commerce platforms under the background of big data.