摘要
针对决策树构造中存在的最优属性选择困难、抗噪声能力差等问题,提出了一种新的基于变精度粗糙集模型的决策树构造算法。该算法采用近似分类精度作为节点选择属性的启发函数,与传统基于粗糙集的决策树构造算法相比,该算法构造的决策树结构简单,提高了决策树的泛化能力,同时对噪声也有一定的抑制能力。
Aiming at the problems of choosing the best attribute and eliminating the influence of the noise data,a new decision tree construction algorithm based on the variable precision rough set model is proposed,which takes the approximate classifying precision as the heuristic function of choosing attributes at a node.Compared to the traditional decision tree classification algorithm based on rough set,the decision tree construct with this method is not only simple-structured,but also antinoise.
出处
《计算机与数字工程》
2013年第3期337-339,共3页
Computer & Digital Engineering
基金
国家高新技术研究发展计划(863计划)项目(编号:2012AA062105)资助
关键词
数据挖掘
决策树
变精度粗糙集
近似分类精度
data mining
decision trees
variable precision rough set
approximate classifying precision
