Magnification Control in Winner Relaxing Neural Gas

摘要

We transfer the idea of winner relaxing learning from the self-organizing map to the neural gas to enable magnification control independently of the shape of the data distribution.