Research of Radar Range Profile´s Recognition Based on an Improved C-SVM Algorithm

This paper improves upon Support Vector Machines (SVM) algorithm on unclassifiable sample sets condition for more to enhance its applicability, which is named after C-SVM (C is a parameter). One hand, non-equidistant margin hyper plane (NM) in high dimension eigen space is introduced to improve on study precision, On the other hand, effectual training sample sets in high dimension eigen space are filtrated, via algorithm introduced by this paper, to reduce study time. Above-mentioned methods are applied to Radar Range Profile´s Recognition, experimental results show that these methods can give very excellent recognition effect.