基于用户偏好及稀疏数据扩充的推荐算法

打开文本图片集

中图分类号：TP391 文献标识码：A 文章编号：2096-4706（2025）09-0107-05

Abstract： To address the data sparsity issue existing in traditional collaborative filtering recommendation algorithms based on user preference， a recommendation algorithm named RA-UPSDE （Recommendation Algorithm based on User Preferences and Sparse Data Enrichment） is proposed. This algorithm firstly improves the traditional calculation of user preference degrees， then enriches the original rating matrix by considering both the number and the method of data enrichment. Finally， it integrates the user preference similarity with the user rating similarity obtained from the enriched data to calculate and perform collaborative filtering recommendations for the target users. Experimental results demonstrate that the proposed algorithm alleviates the data sparsity issue and further enhances the recommendation quality.

Keywords： user preference; sparse data; similarity; data enrichment

0 引 言

随着信息技术的迅猛发展，互联网信息呈现出爆炸性增长，每天新增数据量高达 1018 字节，远超人们的承受范围，因此出现信息过载现象 [1]。（剩余6256字）