Research On Personalized Recommendation Of Dresses Based On Product Key Attributes

With the constant growth and development of the e-commerce platform,there have emerged some problems of information overload while more choices of products for users are provided,and users often have to spend a lot of time and energy to find their expected products.However,with the acceleration of people's pace of life,consumers are increasingly focusing on shopping efficiency.Therefore,the e-commerce platforms have launched personalized recommendation functions in order to improve the shopping efficiency of users.Personalized recommendation service can recommend products to users contrapuntally to help users find their expected products quickly,save their shopping time,improve their shopping experience and improve the marketing conversion rate as well.Apparel is an important commodity category on e-commerce platforms,so it is of great significance to provide users with personalized recommendations for apparel.The product attribute is a component unit of the product.Understanding the consumers' preference for the product attribute is conducive to the effective development and accurate recommendation of products.One-piece dress is an important product in women's apparel.Staring from the analysis of attributes of one-piece dress,this paper conducts a research and data analysis of female consumers' preference of one-piece dress so as to build a personalized recommendation model of one-piece dress based on product key attributes.The specific process of building the model is as follows:(1)To combine with the professional knowledge of apparel to find out the attributes and attribute values of one-piece dress based on the literature review,expert interview and the reference to the classification of one-piece dresses on the e-commerce platforms,and to compile a questionnaire to conduct a survey of consumers' preference of one-piece dress;(2)To obtain the consumers' physical characteristics,levels of consumption,concern about various attributes of one-piece dress when purchasing,preference of various attribute values of one-piece dress and to know the relationship of age and physical characteristics with their preference of various attributes of one-piece dress based on questionnaire,descriptive statistical analysis and correlation analysis;(3)To extract key attributes of one-piece dress the according to the consumers' concern about various attributes of it and to build a one-piece dress model,that is,a product model;(4)Based on the relationship between the consumer attributes and their preference of the attribute of one-piece dress,to get their preference of various attribute values of one-piece dress and select consumers' characteristic attributes and attribute values,and to build a consumer model,that is,a user model;(5)To invite apparel experts to evaluate the importance of the attributes of one-piece dresses of various styles to calculate the weight of the attributes of each style;(6)To build a personalized recommendation model of one-piece dress based on the product key attributes.Based on the personalized recommendation model of one-piece dress,this paper has designedand developed a personalized dress recommendation system.The main function of the system is to provide users with personalized recommendations for one-piece dress products based on season,age,price and physical characteristics.Apparel sellers can also utilize the personalized recommendation function to identify the people who fit the clothes well and train the marketing personnel to improve the efficiency of shopping guide.In order to verify the recommendation effect of personalized recommendation system of one-piece dress,50 styles of one-piece dresses are selected and introduced into the personalized recommendation system of one-piece dress,and the recommendation system is used to recommend for 20 subjects who will evaluate the recommended one-piece dressed which are recommended for them.The experimental result shows that the user satisfaction and prediction accuracy are relatively high,and the recommendation effect of the system is good.