A personalized recommendation method for product copywriting based on user portaits
BU Chunqing
DING Jun
Abstract：In the current e-commerce environment,information overload makes user decision-making increasingly difficult.However,the sparse interaction between users and product copywriting,the dynamic changes in user interests,and insufficient utilization of multimodal information by traditional recommendation methods have constrained the effectiveness of recommendations and user experience.To enhance the effect of product copywriting recommendation and user satisfaction,a personalized product copywriting recommendation method based on user portraits is proposed.By integrating multidimensional data such as user basic attributes,interaction behaviors,and comment feedback,a quantitative user portrait model is constructed.The item-based collaborative filtering algorithm is used to calculate the similarity of copywriting and generate candidate recommendation sets.Based on the similarity between user portraits and copywriting feature vectors,the optimal copywriting is selected for recommendation.Experiments based on e-commerce platform data show that this method can accurately depict user characteristics,with a copywriting coverage rate of 71%to 92%and user satisfaction exceeding 90%.It effectively achieves personalized recommendation of product copywriting and has good practicality and promotional value.
Keywords：user portraitproduct copywritingpreference tendencycollaborative filtering
Publication Date：2025-10-25
Online Publishing Date：2025-11-24(First online date of this platform, not the publication date of the document)
Pages：6( 16-20,30 )