Polygonal Inductive Generalisation System

Classification learning has been dominated by the induction of axisorthogonal decision surfaces. While induction of alternate forms of decision surface has received some attention in the context of decision trees, this issue has received little attention in the context of decision rules. An inductive learning algorithm has been developed which creates arbitrarily shaped concepts. Results from a prototype implementation demonstrate that the approach performs well on target concepts that are not readily represented by long, flat decision surfaces. Keywords : classification learning, oblique decision surfaces, non-axisorthogonal decision surfaces, covering algorithms, hyper-polygonal decision regions. 1 Pre-publication draft of paper accepted for publication in the Proceedings of the Eighth International Conference on Industrial and Engineering Applications of Artificial Intelligence and Expert Systems (IEA/AIE '95 ), pages 587-592. Gordon and Breach Science Publishers 

• Type: Conference
• Year: 1995
• Where: IEAAIE
• Authors: Douglas A. Newlands, Geoffrey I. Webb
• Keywords: Artificial Intelligence, Axisorthogonal Decision Surfaces, Decision Surfaces, IEAAIE 1995, Oblique Decision Surfaces