基于GRI模型和C5．0规则集的图书推荐模式研究

A Study of Recommend Modes in Books Service Based on the GRI Algorithm and C5.0-Based Rule Set

通过应Clementine12．0中关联规则的GRI模型和C5．0推理规则集，对问卷涉及的205位青少年用户基本信息和图书购买信息进行数据建模和分析。研究结果表明，单个用户购买的多类图书之间，具有相同购买特征（如性别、年龄、教育背景等）的用户群和某类图书之间均存在一定的关联规律。在此基础上，提出了基于GRI模型和C5．0规则集的图书推荐模式，即跨类型和相同用户兴趣的图书推荐，试图为图书馆和图书零售市场开展图书推荐服务提供参考。

This paper obtains the data of the personal information and books purchased information through the questionnaire survey （n=205）. Moreover, based on the GRI algorithm and CS.0-based rule set in Clementine 12.0, it discovers a thorough understanding about some significant links among all categories of books purchased, but also botwoen one category of books and simultaneous user group flags, such as gender, age, education background and so on. Finally, it intends to put forward a new recommendation model and its application in library services.