新的决策表约简模型下的一种增量算法  

An Incremental Algorithm for the New Reduction Model of Decision Table

作　　者：蒋思宇[1,2] 

机构地区：[1]华中科技大学计算机学院现代数据库与软件工程实验室 [2]解放军通信指挥学院,武汉430010

出　　处：《计算机工程与应用》2005年第28期21-25,32,共6页Computer Engineering and Applications

基　　金："十五"国家部委项目资助

摘　　要：新的约简模型将满足平均决策强度条件的最简规则集作为最终的约简结果,解决了不相容决策表约简结果不一致的问题。当前迫切需要解决的问题是,面对动态变化的决策表,如何高效地调整决策表的约简结果。文章提出了一种在新的约简模型下增量算法,该算法能根据决策表新增样本与原有的最简规则集的关系,快速推出新决策表的满足平均决策强度条件的最简规则集。The most concise decision rule set that satisfies condition of mean decision power is regarded as final reduction result in the new reduction model,lt is contributing for solving the problem that exists in the reduction results of inconsistent decision table.So it is urgent requirement to increase computed velocity for new reduction result.This paper proposes an incremental algorithm for the new reduction model.This algorithm can rapidly induct new reduction result by analyzing relationship between new example and former reduction result,and it is shown by our experiments this algorithm can acquire good performance.

关 键 词：决策表 约简模型 平均决策强度 增量算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]