Fast Support Vector Classifier (A low complexity alternative to SVM)

A low complexity alterantive to SVM for classification problems (single and multiple class)

Implements a low complexity classifier based on LMS training in
a nonlinearly expanded feature space based on simple RBF units.
The centers of the units are support vectors selected from the
training sample using a simple search algorithm based on novelty
detection.
Relevant papers:
R. Dogaru, “A hardware oriented classifier with simple constructive
training based on support vectors”, in Proceedings of CSCS-16, the
16th Int’l Conference on Control Systems and Computer Science,
May 22 - 26, 2007, Bucharest, Vol.1, pp. 415-418.

Dogaru, R. ; Dogaru, I.,
"An efficient finite precision RBF-M neural network architecture using support vectors"
in Neural Network Applications in Electrical Engineering (NEUREL), 2010 10th Symposium on
Digital Object Identifier: 10.1109/NEUREL.2010.5644089
Publication Year: 2010 , Page(s): 127 - 130