An efficient Self-organizing map learning algorithm with winning frequency of neurons for clustering application

The Self-organizing map (SOM) has been extensively applied to data clustering, image analysis, dimension reduction, and so forth. The conventional SOM does not calculate the winning frequency of each neuron. In this study, we propose a modified SOM which calculate the winning frequency of each neuron. We investigate the behavior of modified SOM in detail. The learning performance is evaluated using the three measurements. We apply modified SOM to various input data set and confirm that modified SOM obtain a more effective map reflecting the distribution state of the input data.