Neurocomputing in Space - High Dimensional Neurocomputing: Growth, Appraisal and Applications - page 121

Fig. 6.2 a Similarity transformation in 3-D; b Scaling and translation in 3-D space

Example 6.2 In this experiment the network is trained for input-output mapping
over a straight line for similarity transformation (scaling factor 1 2) only. The gen-

eralization ability of such trained network is tested over cylinder containing 202 data
points. The transformation result in Fig. 6.2 a shows the excellent generalization with
proposed methodology.

/