A Tool for Model Generation and Knowledge Acquisition

Abstract

Tools to automatically induce domain descriptions fromexamples are valuable aids for the knowledge acquisition stage of expert systemconstruction. This paper presents a description of an algorithm that induces twodomain descriptions: a conceptual model, which gives a broad understanding ofvariable interactions and their effect on the system, and a predictive model,which determines the system value associated with variable values input to themodel. This induction algorithm is based on Entropy Data Analysis (EDA), whichbuilds linear equations to approximate the system described by the trainingdata.