A.I. Approaches in Statistics
Robert Phelps, Peter Musgrove · Brunel University Research Archive (BURA) (Brunel University London) · 1985
The role of pattern recognition and knowledge representation methods from artificial Intelligence with in statistics is considered. Two areas of potential use are identified and one, data exploration, is used to illustrate the possibilities. A method is presented to identify and seperate overlapping groups within cluster analysis, using an A.I. approach. The potential of such 'intelligent ' approaches is stressed. The human brain in its development throug infancy faces a primary task of making sense of its environment. To achieve this it is adapted to extract structure from its environment using both inborn and learned mechanisms.