融合协同过滤与关联规则的图书馆图书选择算法OA

This study proposes a library book selection algorithm integrating collaborative filtering and association rules,aiming to solve the problems of data sparsity and insufficient recommendation diversity of traditional book recommendation methods.The proposed algorithm reasonably fills the missing values in the sparse matrix by combining user similarity and book similari-ty,and uses association rules to mine potential associations between books,so as to improve the accuracy and diversity of rec-ommendations.The experiment is validated by the LibraryThing dataset and the Book-Crossing dataset.The results show that the root mean square errors of the proposed algorithm are 0.89 and 0.84,the mean absolute errors are 0.72 and 0.69,and the average response time is 76.57 ms.Compared with content-based recommendation,matrix factorization and deep neural net-work recommendation algorithms,the proposed algorithm shows significant advantages in accuracy,recommendation diversity and real-time performance,and can provide an effective solution for personalized recommendation in libraries.