Self Organized Maps (SOM) neighborhood and weight updates -
i studying self organized map (som) in neuron nets field. have 2 questions:
1) why neighborhood size decreasing? 2) why not update winner? happen in case?
thanks in advance
the power of som create neural network displayable , human readable, so:
1) neighborhood size decreased in order stability in algorithm increasing iterations.
2) mean of updating neighborhood create map (that displayable) near units have similar weights. if update winner unit, map not created since similar units left scattered in map.
Comments
Post a Comment