Bb. Chaudhuri et U. Bhattacharya, Efficient training and improved performance of multilayer perceptron in pattern classification, NEUROCOMPUT, 34, 2000, pp. 11-27
In pattern recognition problems, the convergence of backpropagation trainin
g algorithm of a multilayer perceptron is slow if the concerned classes hav
e complex decision boundary. To improve the performance, we propose a techn
ique, which at first cleverly picks up samples near the decision boundary w
ithout actually knowing the position of decision boundary. To choose the tr
aining samples, a larger set of data with known class label is considered.
For each datum, its k-neighbours are found. If the datum is near the decisi
on boundary, then all of these k-neighbours would not come from the same cl
ass. A training set, generated using this idea, results in quick and better
convergence of the training algorithm. To get more symmetric neighbours, t
he nearest centroid neighbourhood (Chaudhuri, Pattern Recognition Lett. 17
(1996) 11-17) is used. The performance of the technique has been tested on
synthetic data as well as speech vowel data in two Indian languages. (C) 20
00 Elsevier Science B.V. All rights reserved.