A Parameter in the Learning Rule of SOM That Incorporates Activation Frequency
In the traditional self-organizing map (SOM) the best matching unit (BMU) affects other neurons, through the learning rule, as a function of distance. Here, we propose a new parameter in the learning rule so neurons are not only affected by BMU as a function of distance, but as a function of the frequency of activation from both, the BMU and input vectors, to the affected neurons. This frequency parameter allows non radial neighborhoods and the quality of the formed maps is improved with respect to those formed by traditional SOM, as we show by comparing several error measures and five data sets.
KeywordsInput Vector Learning Rule Error Quantization Frequency Function Learning Factor
Unable to display preview. Download preview PDF.
- 2.Kirk, J., Zurada, J.: A two-stage algorithm for improved topography preservation in self-organizing maps. Int. Con. on Sys., Man and Cyb. 4, 2527–2532 (2000)Google Scholar
- 4.Kohonen, T.: Self-Organizing maps, 3rd edn. Springer, Heidelberg (2000)Google Scholar
- 5.Flanagan, J.: Sufficiente conditions for self-organization in the SOM with a decreasing neighborhood function of any width. Conf. of Art. Neural Networks. Conf. pub. No. 470 (1999)Google Scholar
- 7.Ritter, H.: Self-Organizing Maps on non-euclidean Spaces Kohonen Maps. In: Oja, E., Kaski, S. (eds.), pp. 97–108 (1999)Google Scholar
- 11.Kiviluoto, K.: Topology preservation in Self-Organizing maps. In: Proc. ICNN 1996, IEEE Int. Conf. on Neural Networks (1996)Google Scholar
- 12.Venna, J., Kaski, S.: Neighborhood preservation in nonlinear projection methods: An experimental studyGoogle Scholar