摘要
在介绍了一些典型的决策树分类算法的基础上 ,论述了所开发的面向先进制造企业的综合数据挖掘系统ScopeMiner中的决策树分类算法 ,该算法集中了ID3、C4 5和MedGen等典型算法的思想 ,并进行了改进·在建立决策树过程中采用关联性度量的计算来确定划分条件属性的顺序 ,通过阈值设定和处理简化了决策树的剪枝和优化过程 ,准确性高 ,分类速度快·系统已在某大型企业质量控制中得到了应用 ,取得了一定的经济和社会效益·文章详述了算法的执行过程、应用于冶金企业中的实例以及正确性证明和时间复杂性分析·
A decision tree classification algorithm was described. This algorithm was used in ScopeMineran integrated data mining system for advanced manufacturing enterprise. The algorithm used the idea of ID3,C4 5 and MedGen,and implement some modifications,which included evaluating condition attributes with correlation,and made pruning and optimization process simplify in order to get high accuracy and fast classifying speed. ScopeMiner system has been used in a large enterprise to control production quality and made some economic and society profits. The executing process of the algorithm was expounded. A case of using in a metallurgy enterprise was presented. The accuracy of the algorithm was proven and the complexity of time was analyzed.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2001年第5期481-484,共4页
Journal of Northeastern University(Natural Science)
基金
跨世纪优秀人才培养计划基金资助项目
高等学校骨干教师培养计划基金资助项目
