Local linear correlation analysis with the SOM

The purpose of this paper is to illustrate a method which can be used to select relevant input variables for non-linear regression. The proposed method is an extension to the concept of SOM such that the linear correlation coefficient is computed over a whole data manifold in neighbour subspaces. Using the topographic properties of the usual SOM a localised correlation coefficient may be obtained by modified Kohonen learning. The graphical ordered plot of the obtained local correlation allows to study the non-linear dependencies of variables


Published in:
Neurocomputing, 21, 1-3, 79 - 90
Year:
1998
ISSN:
0925-2312
Keywords:
Note:
local linear correlation analysis;SOM;relevant input variable selection;nonlinear regression;linear correlation coefficient;data manifold;neighbour subspaces;topographic properties;localised correlation coefficient;modified Kohonen learning;graphical ordered plot;nonlinear dependencies;
Laboratories:




 Record created 2007-04-04, last modified 2018-03-18


Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)