一种基于GAAA算法的决策树优化算法

Adecision Tree Optimination Algorithm Based on GAAA Algorithm

决策树是一种比较有效的数据挖掘方法,缺点是当训练集数据属性很多时,构建的决策树的规模会随着属性个数增加而增长。论文从GAAA算法的角度,提出一种动态融合的方法,确定最佳融合时刻。实验结果表明该算法可以有效克服停滞,提高搜索效率,有效地挖掘出最优的分类规则集。

The decision tree is an effective data mining methods.The disadvantage is that when the attributes of training set data are more,the size of the constructed decision tree will grow with the increase of the number of attributes.This paper proposed a dynamic integration from the perspective of GAAA Algorithm to determine the optimum integration time.The experimental results show that the algorithm can effectively overcome stagnation,improve search efficiency and effectively mine the optimal classification rule set.