Rule Bases Integration on the Example of a Fuzzy Model of WIG20.

Sometimes, when a fuzzy model is created, two knowledge sources are available – a data set describing the behavior of an analyzed system and a set of expert rules. In such case three possible approaches to the modelling process are possible: utilize a data set, utilize a set of expert rules or utilize both knowledge sources simultaneously. According to the author of this paper, the third approach is the most reasonable because by joining expert rules with rules created on the basis of a data set, fuzzy model inherits the specific advantages of both rule sets. The aim of this paper is to verify this statement and to illustrate (on a real example) the superiority of a model composed of expert and ˵data″ rules over a model composed only of expert rules and a model composed only of ˵data″ rules. The paper presents a writer΄s method of combining two rule bases and its application for building a combined fuzzy model of a real economic problem. [ABSTRACT FROM AUTHOR]