示例学习的广义扩张矩阵算法及其实现  

A GENERALIZED EXTENSION MATRIX ALGORITHM OF LEARNING FROM EXAMPLES AND ITS IMPLEMENTATION

作　　者：赵美德[1] 李星原[1] 洪家荣[1] 陈彬[1] 

机构地区：[1]哈尔滨工业大学计算机科学与工程系

出　　处：《计算机学报》1994年第9期703-707,共5页Chinese Journal of Computers

摘　　要：本文对扩张矩阵理论加以扩充，提出关于公式的扩张矩阵的概念，并据此实现一个广义扩张矩阵算法叫做ＡＥ９．本文还将ＡＥ９和ＡＱ（１５）应用于几个实际领域的学习问题，如睡眠状态的分类，手写数字识别等．结果都表明ＡＥ９比ＡＱ（１５）分类精度更高．In order to optimize extension matrix algorithm,this paper presents a new version of the extension matrix algorithm called AE9.To make a comparison with AQ15,AE9 is tested for several real-world problems,such as learning the decision rules of human sleep stages,and learning the classification rules of unconstrained handwritten numerals. The results show that AE9 is always more accurate than AQ15.

关 键 词：示例学习 知识获取 扩张矩阵 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]