Integrated Intelligent Decision-Making and Information Discovery.
Brian Xu, Andrew A. Kostrzewski · 2006
Abstract – The Integrated Intelligent Decision-Making and Information Discovery (IIDI) system was created for data management command and control (C2) systems. This paper focuses on one IIDI technique, grid fuzzy neural networks (GFNNs), which can learn and support decision making, and information/knowledge discovery. The IIDI GFNNs were developed, trained, and tested on real sample datasets to detect or discover system attacks and abnormalities on a spacecraft. The IIDI prototypes were developed and demonstrated, including both client-server and Web-based software packages that accurately detected abnormalities and potential attacks in real sample datasets. Our experiments proved that the IIDI GFNNs require less training time, while being two to three times more accurate and 15 to 25 times faster than commercial neural net products. In addition, the IIDI system can be retrained to learn, to discover information, to and perform new tasks in new environments. Keywords: Intelligent system, Neural network, Decision-making, Information technology In recent years, technologies such as standard neural networks (SNNs) and evolutionary algorithms have been applied to pattern recognition, data mining, and knowledge discovery. However, no single technique alone, including SNNs and evolutionary algorithms, can solve real world problems in either industry or the military. For instance, SNNs exhibit slow learning and low accuracy [1]. These weaknesses pervade all commercial SNN