Summary
International Symposium on Nonlinear Theory and its Applications
2017
Session Number:A2L-C
Session:
Number:A2L-C-3
Continuous Learning of the Som with an Adaptive Neighborhood Function
Hikari Yoshimi, Hidetaka Ito, Hiroomi Hikawa,
pp.148-151
Publication Date:2017/12/4
Online ISSN:2188-5079
Summary:
This paper proposes a new neighborhood function for the self-organizing map (SOM). As the learning of the SOM progresses, the conventional neighborhood function shrinks its magnitude and neighborhood radius, and the learning stops after pre-defined training iterations. On the other hand, the proposed neighborhood function uses only the distance between the weight vector of the winner neuron and the input vector, then themagnitude and radius are computed according to this distance. Since the proposed neighborhood function is not a function of the learning iterations, it allows the SOM to continue its learning without time constraints. This feature is especially effective under the changing input vector space that arises in, e.g., online learning.