Intelligent target recognition based on hybrid support vector machine

In this paper we presented a new method of adaptive projective algorithm to improve the speed of SVM classifier. The idea is to pre-extract support vector from training pattern, so that the classification speed is increased with the same performance. Besides this method, the modifying kernel function method is used to achieve high precision approximation. Therefore a novel hybrid support vector machine based on adaptive projective algorithm and modifying kernel functions method for the intelligent target recognition can increase the generalization ability and the speed. It is shown that the radar target data can be classified, and consequently the separability between classes is increased and speed is well improved with this hybrid support vector machine.