Title :
A software development process model for artificial neural networks in critical applications
Author :
Rodvold, David M.
Author_Institution :
Colorado Tech. Univ., CO, USA
Abstract :
Using artificial neural networks (ANNs) in critical applications can be challenging due to the often experimental nature of ANN construction and the “black box” label that is frequently attached to ANNs. Well-accepted process models exist for algorithmic software development which facilitate software validation and acceptance. The software development process model presented herein is targeted specifically toward ANNs in critical applications. This model is not unwieldy, and could easily be used on projects without critical aspects. This should be of particular interest to organizations that use ANNs and need to maintain or achieve a capability maturity model or ISO software development rating. Further, while this model is aimed directly at neural network development, with minor modifications, the model could be applied to any technique wherein knowledge is extracted from existing data, such as other numeric approaches or knowledge-based systems
Keywords :
digital simulation; neural nets; performance evaluation; software engineering; critical applications; learning; nested loop model; neural network development; risk management; software development process model; Application software; Artificial neural networks; Capability maturity model; Computer networks; Intelligent networks; Neural networks; Power system modeling; Programming; Risk management; Software algorithms;
Conference_Titel :
Neural Networks, 1999. IJCNN '99. International Joint Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5529-6
DOI :
10.1109/IJCNN.1999.836192