Year
1994
Abstract
In fuzzy control, expert knowledge is encoded in the form of fuzzy rules, which describe recommended actions for different classes of situations represented by fuzzy sets. The concept of a fuzzy controller is applied to the forecasting problem in a time series, specifically, to forecasting and detecting anomalies in inventory differences. This paper reviews the basic notion underlying the fuzzy control systems and provides examples of application. The well-known materialunaccounted- for diffusion plant data of Jaech are analyzed using both feedforward neural networks and fuzzy controllers. By forming a difference between the forecasted and observed signals, an efficient method to detect small signals in background noise is implemented.