论文部分内容阅读
[目的/意义]为解决高校图书推荐过程中面临的“数据稀疏”和“冷启动”问题,研究表明:优化读者评价矩阵和相似度模型是提高图书推荐质量的关键。[方法/过程]提出一种协同过滤改进方法,以图书分类为项目生成用户评价矩阵,并考虑借阅方式、借阅时间和图书相似度对用户兴趣度的影响,优化矩阵中的样本数据;同时,在计算读者相似度时融入读者特征和图书特征。[结果/结论]实验结果表明,该方法可有效解决“数据稀疏”和“冷启动”问题,显著降低计算量。与基本协同过滤和聚类改进的协同过滤方法相比,无论是在推荐准确率还是在用户满意率上都有较大的提高,综合推荐效果更好。
[Objective / Significance] To solve the problems of “data sparse” and “cold start” in the process of university book recommendation, the research shows that optimizing reader evaluation matrix and similarity model is the key to improve the quality of book recommendation. [Methods / Processes] This paper proposes a collaborative filtering improvement method, which uses books as a project to generate user evaluation matrix and optimizes the sample data in the matrix by considering the influence of borrowing methods, borrowing time and book similarity on the user’s interest. At the same time, Readership and book features are included when calculating reader similarity. [Result / Conclusion] The experimental results show that the proposed method can effectively solve the problems of “data sparse” and “cold start”, and significantly reduce the computational cost. Compared with the collaborative filtering method based on basic collaborative filtering and clustering improvement, the comprehensive recommendation is better in both the recommendation accuracy rate and the user satisfaction rate.