基于贝叶斯粗糙集和混合专家模型的CBR系统  

Bayesian rough set and mixture experts model based CBR system

作　　者：韩敏[1] 王心哲[1] 李洋[1] 童年 

机构地区：[1]大连理工大学电子信息与电气工程学部,辽宁大连116023 [2]辽宁金自天正智能控制股份有限公司,沈阳110010

出　　处：《控制与决策》2013年第1期157-160,共4页Control and Decision

基　　金：国家自然科学基金项目(60674073);国家863计划项目(2007AA04Z158)

摘　　要：建立一个完整的案例推理系统,提出一种高效的案例检索方法和一种案例调整策略.在案例检索过程中,提出一种基于贝叶斯粗糙集的属性权重确定算法,在此基础上利用最邻近法检索出与当前案例最相似的一组案例作为参考.使用检索出的相似案例训练分层混合专家模型,并用微粒群算法优化模型参数,实现了对案例的调整.采用实际转炉生产数据进行仿真,结果表明了该案例推理系统的有效性.An integrated case-based reasoning system is built, and an efficient case retrieval method and a case adjust method are proposed in the system. In the case retrieval process, an algorithm to compute attributes weight is presented based on improved Bayesian rough set model. Then, the most nearest neighbor method is used to retrieve a group of similar cases as the reference of current furnace. In case adjust process, a hierarchical mixture of experts model is trained with this group of similar cases, then particle swarm optimization algorithm is adopted to optimize the parameters. A simulation experiment is implement with practical data from a steel plant. The results show the effectiveness of the proposed case-based reasoning system.

关 键 词：案例推理 粗糙集 分层混合专家模型 微粒群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]