Automatic learning parameters for self-organizing feature maps

This paper presents a method which automatically determines the learning parameters of a self organizing feature map during the learning. Therefore, system models of the learning and organizing process are developed in order to be followed and predicted by linear and extended Kalman filters. The Kalman filters estimate the learning parameters optimal within the system models, so that the self organizing process converges automatically to a neighbourhood preserving feature map of the learning data. Finally, the estimation method is demonstrated using data from linear and nonlinear manifolds