Hyperparameter Selection for Self-Organizing Maps

Akio Utsugi
1997 Neural Computation  
A topology-selection method for self-organizing maps (SOMs) based on empirical Bayesian inference is presented. This method is natural extension of the hyperparameter-selection method presented earlier, in which the SOM algorithm is regarded as an estimation algorithm for a Gaussian mixture model with a Gaussian smoothing prior on the centroid parameters, and optimal hyperparameters are obtained by maximizing their evidence. In the present paper, comparisons between models with different
more » ... ies are made possible by further specifying the prior of the centroid parameters with an additional hyperparameter. In addition, a fast hyperparametersearch algorithm using the derivatives of evidence is presented. The validity of the methods presented is confirmed by simulation experiments.
doi:10.1162/neco.1997.9.3.623 fatcat:mb45yh56lrf35ighsday4tylli