基于协同过滤和特征工程的图书推荐系统研究及云图构建

Research on Book Recommendation System and Cloud Visualization Construction Based on Collaborative Filtering and Feature Engineering

面向图书馆联机公共目录检索(OPAC)系统,提出一种基于协同过滤结合特征工程的加权图书推荐算法。所提算法运用协同过滤的思想,通过分析图书借阅历史,对兴趣相似的用户进行聚类。引入图书借阅次数、上架时间、排架相邻度、专业相关度等特征项,经过特征编码后,利用加权算法计算出图书推荐指数。此外,针对推荐系统的冷启动和稀疏性问题,提出解决策略。测试结果表明,所提算法实现了完整的图书推荐系统及推荐图书的云图构建。

This paper proposes a weighted book recommendation algorithm for online public access catalogues(OPAC)system,by combining collaborative filtering with feature engineering.The algorithm applies collaborative filtering to cluster users with similar interests by analyzing the borrowing history.The book recommendation index is calculated by weighted algorithm after the feature code is introduced,such as the number of books borrowed,shelf time,shelving adjacency and professional relevance.In addition,the solutions to the problems of cold start and sparsity are proposed.The test results show that the proposed algorithm realizes the complete book recommendation system and the cloud visualization construction of recommended books.