【作者】 刘晋； 王潜平； 房坤；

【Author】 Liu Jin Wang QianPing Fang Kun ( College of mining and technology, department of computer science and technology, XuZhou 221008)

【机构】 中国矿业大学计算机科学与技术学院；

【摘要】 本文从基于近邻用户的协作过滤思想出发,提出一种新的个性化推荐方法,该方法通过为用户创建SIGAgent来模拟用户,挖掘用户的点击情况并形成反馈,使兴趣相似的用户自动聚集并相互推荐,并利用反馈信息对系统进行调整。同其他个性化推荐方法相比,本方法简单直观,更适合大规模用户和资源情况下的个性化推荐。文中详细描述了系统的结构和流程,并对模拟用户的SIGAgent和系统的工作原理进行了重点描述。更多还原

【Abstract】 This paper starts from the idea of collaborative filtering and k-nearest neighbor method, proposes a new way to make personalized recommendation. This approach simulates users by creating SIGAgents accordingly, digs clicks of users and forms feedbacks, making users with similar interests assemble automatically and recommend to each other, and makes necessary adjustments to feedbacks. Compared with other personalize recommend approaches, this is much a more simple and intuitivism way, especially in condition with large amount of users and resources. Architecture and process of the system is explicated and structure of SIGAgent and principles of the system are illustrated in detail.更多还原