A Knowledge Roughness Based Approach to Hybrid Decision Tree Construction
Hongmei Lu, Xuegang Hu, Baosheng Yang · Procedia Engineering · 2012
It is difficult for unvaried decision tree to reflect the relationship of attributes, multivariate decision tree can resolve this problem preferably, the former produces big tree, the latter gains simple tree but difficult to explain. Aim to these two features for decision tree, in this paper we advance a knowledge roughness based approach to hybrid decision tree, select lesser knowledge roughness as tested attribute to construct decision tree. As a result,we find this is a good approach with simple operation and higher efficiency.