Data-driven fuzzy rule induction and its application to systems monitoring

This paper presents a novel approach to the development of a modular technique for inducing fuzzy rule sets. It allows for automated, data-driven generation of sets of fuzzy decision rules from historical data, thereby facilitating faster knowledge acquisition for knowledge-based applications. A rough set-based preprocessor is employed to reduce the dimensionality of complex data sets. The facility of this technique is demonstrated by applying it to building a monitoring system for a large urban water treatment plant.