Abstract: 【Purpose/significance】In the process of social tagging,tag frequency can not accurately represent user preferences,as well as the problem of data sparsity and cold start in the process of book recommendation【. Method/process】Based on the memory optimiza⁃tion mechanism,this paper proposes a method to express readers' preferences.Taking "Douban Shudu" as the empirical object,the DB⁃SCAN clustering results are used to evaluate the method.and the experimental results show that the method has good representation ef⁃fect.In order to solve the problems of cold start and data sparsity in the process of book recommendation,this paper studies book group
recommendation on the basis of reader preference representation based on memory optimization mechanism【. Results/conclusion】The experimental results show that the proposed method has high accuracy,recall and F value【. Innovation/limitation】This paper proposes a method to express readers' preferences based on memory optimization mechanism, which is of great significance to mining readers'preferences and developing recommendation services.However,in the process of constructing readers' preferences,we need to further refine the process of cognitive construction and updating.If necessary,we can consider using more readers' attributes to improve the mechanism of memory construction and optimization.