决策树分类技术研究

The Study on Decision Tree Classification Techniques

决策树分类是一种重要的数据分类技术。ID3、C4.5和EC4.5是建立决策树的常用算法，但目前国内对一些新的决策树分类算法研究较少。为此，在消化大量文献资料的基础上，研究了CART、SLIQ、SPRINT、PUBLIC等新算法，对各种决策树分类算法的基本思想进行阐述，并分析比较了各种算法的主要特性，为数据分类研究者提供借鉴。

Decision tree is one of the most important data classification techniques. Algorithms ID3, C4.5 and EC4.5 are widely used to construct decision trees, but little work on new decision tree algorithms has been done. Some new decision tree algorithms are studied, including CART, SLIQ, SPRINT and PUBLIC. The basic ideas and main features of these algorithms are discussed in this paper.