Functional attribute relations by data Modeling and application for knowledge discovery

Model based knowledge discovery is considered. The data model is established from a data set by learning. For the functional attribute relation determination a novel method is developed. The method consists of repeated application of the data set to the data model where one of the independent variables at the input is stepwise incremented from an initial starting attribute value while the rest of the data set and the model remains to be the same. By means of this process, the variation of each dependent variable at the output of the model is established as a function of each independent variable at the input. For the methods and techniques used, data mining is central to their developments.