Thesis Details

Thesis Title: Automating the knowledge acquisition phase of expert system development using neural networks and induction techniques
Thesis Author: Lee Forrest
Abstract: Expert system are computer programs designed to store the knowledge of an expert in a form non-experts can use to solve problems. Expert systems provide a means of storing information than can be lost due to illnesses, retirement or termination. Expert system have been proven useful in many applications; however, they are difficult to create. The most difficult pha~e of system development is knowledge acquisition. Knowledge acquisition i~; the process of recording the knowledge of the expert in a form the computer can use. Objective induction, subjective induction and neural network technology were used to automate the knowledge acquisition phase of expert system development. Four applications \;ere chosen involving the carding, drawing, and spinning mechanical maintenance and yarn grade prediction. The technologies were ab~e to significantly reduce application time and provided the opportunity for creating expert sy~tems that could learn from raw process data without any programming.