一种新的归纳学习算法——基于特征可分性的归纳学习算法

A NEW INDUCTIVE LEARNING ALGORITHM——SEPARABILITY-BASED INDUCTIVE LEARNING ALGORITHM

本文提出了一种新的基于特征可分性的归纳学习算法(SBI)。与现有各种归纳学习算法相比,该方法直接从特征对不同类型的可分性出发,建立可分性判据,然后形成决策树,可对多种概念进行判别。SBI算法具有直观且计算简便等优点。本文以实例表明了SBI算法的有效性。

A new separability-based inductive learning algorithm is proposed in this paper. The algorithm is different from existing inductive learning algorithms. Starting directly from the separability of features for different classes, building a separability criterion, then forming a decision tree, the algorithm can classify multiclass concepts. The algorithm is intuitive, simple, and convenient for computation. Its effectineness is illustrated by an example in this paper.