Research On Optimization Of Personalized Recommendation Combining Product Images And Inventory

How to fully mine the information contained in the product image and construct a prediction method of user needs is an important issue in the research of personalized recommendation systems.Inventory,as an influential factor in a user's shopping process,directly determines whether the user can purchase the desired product.If the limitation of product inventory is ignored during the personalized recommendation process,users will not be able to purchase the recommended products,thereby reducing user satisfaction.Therefore,on the basis of predicting user personalized preferences based on product images,this paper integrates product inventory information to optimize product recommendation results,which helps to improve the accuracy of personalized recommendations and improve user satisfaction.For the above problems,this paper proposes a personalized recommendation optimization method that combines product images and inventory.First,based on the product image data,the deep residual network and the Kmeans method are used to label the products,fully mine the user preference information contained in the product images,establish the relationship between the products,and reduce the impact of data sparsity on the recommendation accuracy.Secondly,construct a user preference prediction model based on the feature division dirichlet process to analyze the user's multi-dimensional interests.Finally,build an optimization strategy for recommendation results based on product inventory to generate a personalized recommendation list.The experimental results on the Taobao data set show that the personalized recommendation optimization method based on product images and inventory proposed in this paper can effectively improve the performance of the recommendation system.The research in this paper proposes effective solutions for improving the effectiveness of personalized recommendation systems by making full use of information such as product images and inventory,provides new ideas for improving user satisfaction in e-commerce environments,and has theoretical and practical significance for the research of personalized recommendation system.