Does the shape of the neuron lattice influence the performance in a Self Organising Map?

by R.p.T   Last Updated February 01, 2018 16:19 PM

I was reading this paper [1] about Self Organising maps, where I came across this particular passage

The neurons are organized in a 2-dimensional map. The ratio of the side lengths of the map is approximately the ratio of the two largest eigenvalues of the training data’s covariance matrix.

Does the ratio of the side lengths really influence the performance of the SOM?

Anomaly detection using SOM and KNN



Answers 1


Shut up, R.p.T. Read first, then ask questions.

Dan
Dan
February 01, 2018 15:45 PM

Related Questions


Clustering Self Organising map nodes

Updated May 28, 2015 10:08 AM

SOM grid size suggested by Vesanto

Updated May 29, 2017 09:19 AM

Empty nodes when creating SOM

Updated March 20, 2017 22:19 PM


Why is SOM not being used?

Updated July 13, 2015 13:08 PM