NEUROFUZZY COGNITIVE MAPS WITH ADAPTIVE B-SPLINES

Cognitive maps and time space maps make use of multidimensional scalin g (MDS) techniques to analyze data relating to spatial and environment al preferences and perception. Perceptual configuration of the points is represented by a cognitive map with surface feature interpolation. In this paper we propose three procedures of MDS using Neurofuzzy adap tive modelling with B-splines for surface feature interpolation. The p rocedures are based on: (1) Hayashi's quantifying method of paired com parisons, (2) Torgerson's metrical MDS procedure and (3) Gradient desc ent method which is the basic learning method of adaptive systems such as the artificial neural networks and neurofuzzy modelling. In numeri cal examples, the resultant maps are compared and an application to so ciometry analysis is presented. (C) 1998 Elsevier Science Inc. All rig hts reserved.