论文部分内容阅读
针对目前传统数字图书馆无法为用户提供准确个性的图书推荐服务的问题,提出构建基于协同过滤的图书智能推荐系统。首先对图书进行聚类,构建无缺失的图书评价矩阵,在此基础上根据读者对相似图书的评分预测读者的兴趣爱好,为读者提供个性化的图书推荐。该方法在评分数据极端稀疏的情况下也可以为读者作出准确的图书推荐。最后通过实验验证该推荐方法的有效性和实用性。
Aiming at the problem that the traditional digital library can not provide accurate and personalized book recommendation service to users, this paper proposes to build a book-based intelligent recommendation system based on collaborative filtering. Firstly, the books are clustered to construct a missing evaluation matrix of books, on the basis of which readers ’preferences and interests are predicted according to the readers’ ratings of similar books, and the readers are provided with personalized recommendation of books. This method can also make accurate book recommendations for readers when the scoring data is extremely sparse. Finally, the validity and practicability of the proposed method are verified by experiments.