For the Weka Self Organising Map node, is there a quick way to visualize the corresponding u-matrix?
Alternatively, any other ways to visualize the output?
BTW, input data is chemical coumpounds with RDKit fingerprints, so a chemistry aware visualization would be good.
Precisely my question... I have a RDkit generated fingerprint and a variety of physicochemical characteristics, can input everything into Weka SOM... can't get it to make a nice looking u-matrix.
R integration could save your day, I've once stumbled over the package (som) that does this...
I usually end up resorting to [R] or Python whenever build-in nodes don't quite do what I want.
Another example is the Randon Forest nodes not giving a "proper" variable importance measure...
Hi got the same question here. I have fingerprints of molecule but not sure how to generate SOM and visualize it. Any details on using R or any methods? Thanks!
My R code used the kohonen package.
Quite easy to put into the R-scripting nodes, plot either using the package's built-in plot function or plot manually via ggplot2.