Batch mode and learning rate

[jtpesone]

If somoclu always uses the batch training mode, how is the learning rate used? If the update is done according to the batch training equation given in Wittek et al, 2017 (Somoclu: An Efficient Parallel Library for Self-Organizing Maps), learning rate is not used at all.

[xgdgsc]

You can search for "scale" in code to see.

[jtpesone]

If you cannot provide this as an equation, I will try to dig into the code. It would be good to document it, however. Would be important for the users to know which equations the sw uses.

[xgdgsc]

https://stats.stackexchange.com/a/402360 ?