COMPUTERIZED LEARNING MACHINES APPLIED TO CHEMICAL PROBLEMS - INVESTIGATION OF CONVERGENCE RATE AND PREDICTIVE ABILITY OF ADAPTIVE BINARY PATTERN CLASSIFIERS

被引：83

作者：

JURS, PC

KOWALSKI, BR

ISENHOUR, TL

REILLEY, CN

机构：

[1] Department of Chemistry, University of Washington, Seattle

[2] Department of Chemistry, University of North Carolina, N.C. 27515, Chapel Hill

来源：

ANALYTICAL CHEMISTRY | 1969年 / 41卷 / 06期

关键词：

D O I：

10.1021/ac60275a025

中图分类号：

O65 [分析化学];

学科分类号：

070302 ; 081704 ;

摘要：

Predictive ability and convergence rate of adaptive binary pattern classifiers have been investigated. The effects of the character of pattern sets, categories, and training sets as well as feedback methods, reduction of parameters, inclusion of atypical data, and choice of initial weight vectors are presented. Examples drawn from mass spectrometry are used to demonstrate the behavior of the systems discussed. © 1969, American Chemical Society. All rights reserved.

引用

页码：690 / &