Share Email Print
cover

Proceedings Paper

Neural network programming using rapid application development techniques
Author(s): Garrett S. Harris; Bruce E. Segee; Vincent M. Allen
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Neural networks and fuzzy logic have merged as useful tools for the calibration of arrays of thin film gas senor. The choice of network parameters is essential for acceptable network performance. Often, choosing network-operating parameters involves a search of many possible candidate networks. When the neural network code is incorporated with other code, such as for data capture, presentation, and control, it is often the case that interdependencies are formed between these code segments. Frequently, enhancements, modifications, and fixes to the code lead to an extensive and time-consuming rewrite of many parts of the software. Thus, the need arise for neural network software modules that can be easily incorporated in application software but whose interface is well defined and whose implementation is entirely separate from the functionality it provides. By providing debugged and proven software modules encapsulating neural network functionality that can be simply inserted into any application, the entire software system can be modularized. These modules can be reused easily, and changing the neural network operating parameters no longer involves a complete software rewrite or even a recompile.

Paper Details

Date Published: 23 November 1999
PDF: 8 pages
Proc. SPIE 3856, Internal Standardization and Calibration Architectures for Chemical Sensors, (23 November 1999); doi: 10.1117/12.371290
Show Author Affiliations
Garrett S. Harris, Univ. of Maine (United States)
Bruce E. Segee, Univ. of Maine (United States)
Vincent M. Allen, Modicon Corp. (United States)


Published in SPIE Proceedings Vol. 3856:
Internal Standardization and Calibration Architectures for Chemical Sensors
Ronald E. Shaffer; Radislav A. Potyrailo, Editor(s)

© SPIE. Terms of Use
Back to Top