I have implemented the algorithm to train self organizing maps in Python and it seems to be working well. I checked with some labeled data and the maps are learning the topology well. Here are some images:

Iris Dataset

A text dataset

I am calculating a Unified Distance Matrix for visualization which does not seem to be working well as seen in the 2nd image.

I believe that similarity coloring as seen in the world poverty map developed by Prof. Kaski and Prof. Kohonen would be apt for my application but I am unable to find a clear explanation of how it is to be calculated. The procedure has been outlined in the paper "Tips for Processing and Color-coding of Self-Organizing Maps ", however, it omits some technical details which are necessary to implement the similarity coloring. Here is the image that I am referring to: World Poverty Map

Could someone help me with the exact steps to calculate the similarity coloring map? I have searched everywhere for a month and I am unable to find an explanation.


1 Answer 1


I don't know exactly how the above was created, but I did something similar in the late '90s and reproduced it recently using somoclu in Python. While I've not applied the colours back to the SOM cells, this would be pretty trivial.

I've put the notebook on kyso, I hope it's useful.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.