Using a mahalanobis-like distance to train radial basis neural networks

Authors:
J. M. Valls;R. Aler;O. Fernández
Affiliations:
Computer Science Department, Carlos III University, Leganés (Madrid), Spain;Computer Science Department, Carlos III University, Leganés (Madrid), Spain;Computer Science Department, Carlos III University, Leganés (Madrid), Spain
Venue:
IWANN'05 Proceedings of the 8th international conference on Artificial Neural Networks: computational Intelligence and Bioinspired Systems
Year:
2005

Citing 5
Cited 3

Approximation and radial-basis-function networks

Neural Computation
A sequential learning scheme for function approximation using minimal radial basis function neural networks

Neural Computation
Locally Weighted Learning

Artificial Intelligence Review - Special issue on lazy learning
An efficient MDL-based construction of RBF networks

Neural Networks
Fast learning in networks of locally-tuned processing units

Neural Computation

Input selection for radial basis function networks by constrained optimization

ICANN'07 Proceedings of the 17th international conference on Artificial neural networks
A radial basis function redesigned for predicting a welding process

MICAI'10 Proceedings of the 9th Mexican international conference on Artificial intelligence conference on Advances in soft computing: Part II
Analysis and evaluation in a welding process applying a Redesigned Radial Basis Function

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

Radial Basis Neural Networks (RBNN) can approximate any regular function and have a faster training phase than other similar neural networks. However, the activation of each neuron depends on the euclidean distance between a pattern and the neuron center. Therefore, the activation function is symmetrical and all attributes are considered equally relevant. This could be solved by altering the metric used in the activation function (i.e. using non-symmetrical metrics). The Mahalanobis distance is such a metric, that takes into account the variability of the attributes and their correlations. However, this distance is computed directly from the variance-covariance matrix and does not consider the accuracy of the learning algorithm. In this paper, we propose to use a generalized euclidean metric, following the Mahalanobis structure, but evolved by a Genetic Algorithm (GA). This GA searches for the distance matrix that minimizes the error produced by a fixed RBNN. Our approach has been tested on two domains and positive results have been observed in both cases.