sigma
now has a different decay function and it can be selected via the parametersigma_decay_function
- the user is now warned when
sigma
is higher thansqrt(x^2+y^2)
- a selection of decay functions for both sigma and learning rate has been implemented by @BrandonGarciaWx
- the train method now accept
fixed_points
so that the algorithm can be force to train specific neurons for specific samples