摘要
基于变精度粗糙集模型,对文献[3]提出的生成决策树方法进行改进,把变精度加权平均粗糙度作为属性选择标准,提出一种构造决策树新算法。新算法用变精度近似精度来代替近似精度,能有效地克服噪声数据在构造决策树过程中对刻画精度的影响,使生成的决策树复杂性降低,泛化能力更强。
Based on Variable Precision Rough Sets Model, the decision tree reducing approach presented in Reference [3] is improved. The article presents a new algorithm for constructing decision tree with variable precision weighted mean roughness as the criteria for selecting attribute. The new algorithm effectively overcomes the influence of the noise data in structuring decision tree, reduces the complexity of decision tree and strengthens its extensive ability.
出处
《广西科学院学报》
2007年第2期76-79,共4页
Journal of Guangxi Academy of Sciences
关键词
决策树
粗糙集
变精度
decision tree, rough sets, variable precision
