基于CBR和MADM的多Agent推荐系统

摘要

To resolve the problem of the CBR recommender systems only consider the situation that the product property values are either all accurate or inaccurate, a new recommender system architecture is proposed based on multi-agent and MADM. A hybrid data similarity measure algorithm and TOPSIS multi-attribute decision making method based on distance are designed taking into account the property classification. The function, structure and workflow of the system are designed. A simple simulation is presented to show process of the CBR and MADM approach in the recommender system and to prove that the system is practical.%为解决目前基于CBR的推荐系统只考虑属性值全部为精确或全部为非精确数据的情况,提出一种基于MADM的多Agent推荐系统框架.在考虑了属性分类的基础上设计了基于距离的混合数据类型的相似性度量算法及TOPSIS多属性决策方法,设计了该系统各组成部分功能、结构和流程.模拟算例演示了案例推理及多属性决策在本系统的应用过程,结果表明该系统有较好的实用性.