CAQ: Making AQ Work with Engineering Problems
Bradley L. Whitehall, Robert E. Stepp · Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign) · 1989
COSATI COOES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)inductive learning systems, learning algorithms, CAQ (Continuous AQ), AQ learning algorithm Engineers need learning algorithms that can work with continuous (real) variables and with nominal values.It i-s not sufficient for an algorithm to only handle one class of variable types.Engineering problems also demand that these methods be able to handle noise and produce the concepts in a comprehensible manner.CAQ (Continuous AQ) is an algorithm designed to meet these requirements.This report describes the algorithm and why combining continuous and nominal values is important.It is shown that handling continuous variables as real numbers instead of forcing them into a nominal representation leads to more efficient processing.22a.