我国学术数据库协同过滤资源推荐研究现状探讨

The Literature Review of Collaborative Filtering Recommendation in Chinese Academic Databases

通过分析2003年至2016年中国知网发表的"图书情报与数字图书馆"领域关于协同过滤推荐的82篇文献,总结我国学术数据库协同过滤资源推荐的研究现状。通过对文献样本进行分类,发现目前该领域的研究重点主要集中于对学术数据库协同过滤推荐的推广和对推荐算法本身的完善两个方面,且后者主要集中于对数据稀疏性问题和可扩展性问题的解决。通过进一步分析,发现国内研究人员主要通过结合基于内容的推荐、空值填补和推荐结果融合三种方法缓解数据稀疏性问题;通过聚类的方法缓解可扩展性问题。

By studying 82 papers published in the CNKI from 2003 to 2016 on collaborative filtering of Information and Digital Library, we investigate the hottest topics of current and history on collaborative filtering of Information and Digital Library. By classifying these papers, we discover that domestic researchers are engaged in either propagating the collaborative filtering technique in academic databases or improving the algorithm, in which the main problems are data sparseness and extensibility. By analyzing the papers further, we discover that researchers prefer to combine the content-based recommend technique, custom data or different recommend results to solve the data sparseness problem, and turn to the clustering technique to solve the problem of extensibility.