基于混合规则的图书推荐模型设计与研究

Books Recommended Model Design and Research Based on Mixing Rules

在海量的数字图书信息中,准确、迅速地找到符合自身需要的图书是数字图书馆服务系统需要解决的主要问题。主要针对数字图书馆服务推荐系统中常采用的单一关联规则或协同过滤规则中存在的不足,构建了一种基于混合规则的数字图书馆服务推荐模型并对传统的推荐算法进行了改进设计。在图书推荐过程中,首先采用关联规则推荐得到若干个用户喜欢的图书子类,然后在这若干个图书子类所属的范围内利用协同过滤规则推荐出用户具体所需的图书。结果表明,该模型不仅能处理图书借阅服务中种类繁多的推荐问题,又可提高推荐效率,具有较高的实用价值。

In the deluge of information in digital books,accurately and quickly find the books meet their needs is a digital library service system needs to solve the main problem.Aimming at the shortcomings existed in the single association rules or collaborative filtering rules used in recommendation system for digital library services,build a hybrid rule-based recommendation model of digital library services and traditional recommendation algorithm is improved.In the book recommendation process,first get a number of sub-class of the book users like recommended by association rules,and then recommend the specific books users require using collaborative filtering rules in a number of sub-class of the book.The results show that the model can not only handle a wide variety of book lending service,recommended the issue,but also improve the efficiency of recommendation,have a high practical value.