Knowledge Discovery and Model Structure Selection Based on Rough Set Theory
-
摘要: 为了合理选择决策模型,提出了一种基于粗糙集理论获取模型选择知识并利用获取的知识选择模型的方法.通过随机设定模型参数,得到备选模型的属性值为连续值的决策表;通过设定误差范围,获得决策表中对象之间的关系;利用该关系和粗糙集理论,对决策表进行属性约简,获得决策规则;按该决策规则进行推理,实现智能决策的模型选择.实例表明,该算法是可行的.Abstract: To select a proper decision model in intelligent decision,a method to acquire the knowledge of model selection and select a decision model by utilizing the knowledge was proposed based on the rough set theory.With this method,the decision table of models with continuous attribute values is acquired by setting parameters in the models randomly.Then,the relation between objects in the decision table is obtained by setting errors.From the relation,the reduction of attributes for the decision table is conducted by applying the rough set theory to gain decision rules.Based on these decision rules,a proper model can be selected. The feasibility of the proposed method has been illustrated using an example.
-
Key words:
- IDSS(intelligent decision support system) /
- rough set /
- model selection /
- decision rule
-
卢涛,李一军,黄梯云.基于类自然语言的管理问题理解方法研究[J]. 管理科学学报,2003,6(1):37-43.LU Tao,LI yijun,HUANG Tiyun.Research on method of managerial problem understanding based on natural language-like problem description language[J]. Journal of Management Sciences in China,2003,6(1):37-43.[2] 黄梯云.智能决策支持系统[M]. 北京:电子工业出版社,2001:119-126.[3] 冯玉强,潘启树.基于神经网络与专家系统结合的模型自动选择[J]. 哈尔滨工业大学学报,2001,33(1):24-27.FENG Yuqiang,PAN Qishu.Automation selection of model based on integration of expert system and neural network[J]. Journal of Harbin Institute of Technology,2001,33(1):24-27.[4] 蓝红兵,费奇.一种基于非单调逻辑的模型管理方法[J]. 自动化学报,1992,18(4):414-420.LAN Hongbing,FEI Qi.A nonmonotonic logic-based method for model management[J]. Acta Automatlca Sinica,1992,18(4):414-420.[5] 张文修,吴伟志,梁吉业,等.粗糙集理论与方法[M]. 北京:科学出版社,2001:26-39.[6] 张文修,梁怡,吴伟志.信息系统与知识发现[M]. 北京:科学出版社,2003:68-75.[7] 刘清.Rough Set与Rough推理[M]. 北京:科学出版社,2001:40-75.[8] QIN Keyun,PEI Zheng.On the topological properties of fuzzy-rough sets[J]. Fuzzy Sets and Systems,2005,151(3):601-613.
点击查看大图
计量
- 文章访问数: 1548
- HTML全文浏览量: 65
- PDF下载量: 256
- 被引次数: 0