Input variable space reduction using dimensional analysis for artificial neural network modeling

被引：0

作者：

Watson, PM ^{[1
]}

Mah, MY ^{[1
]}

Liou, LL ^{[1
]}

机构：

[1] USAF, Res Lab, Sensors Directorate, Wright Patterson AFB, OH 45433 USA

来源：

1999 IEEE MTT-S INTERNATIONAL MICROWAVE SYMPOSIUM DIGEST, VOLS 1-4 | 1999年

关键词：

D O I：

暂无

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

Input variable space reduction using dimentional analysis for artificial neural network modeling of passive components is demonstrated. Results show that using dimensional analysis to limit the number of input variables significantly reduces the amount of training vectors needed for model development, which in turn decreases model development time. Also, dimensional analysis allows for determination of appropriate input variable space and leads to increased model accuracy.

引用

页码：269 / 272

页数：4

共 5 条

[1]

Creech G. L., 1997, IEEE Transactions on Microwave Theory and Techniques, V45, P794, DOI 10.1109/22.575602

[2] Design methodology of microstrip lines using dimensional analysis [J].

Mah, MY ;

Liou, LL ;

Ewing, RL ;

Ferendeci, AM .

IEEE MICROWAVE AND GUIDED WAVE LETTERS, 1998, 8 (07) :248-250

[3]

MIDDENDORF WH, 1990, DESIGN DEVICES SYSTE

[4]

Wang F, 1997, IEEE MTT-S, P627, DOI 10.1109/MWSYM.1997.602870

[5] Design and optimization of CPW circuits using EM-ANN models for CPW components [J].

Watson, PM ;

Gupta, KC .

IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, 1997, 45 (12) :2515-2523

← 1 →