Support Vector Machine (SVM) Based on Membrane Computing Optimization and the Application for C-band Radio Abnormal Signal Identification

Abstract

The Support Vector Machine (SVM) is a widely used tool in classification problems, but the classification performance of Support Vector Machine (SVM) largely depends on the choice of its relevant parameters. This paper proposes a model of Support Vector Machine (SVM) classification based on Cell-like Membrane computing Optimization algorithm (CMO-SVM). In the model, the parameters of Support Vector Machine (SVM) (cost parameter C and RBF kernel parameter ) are optimized by cell-like membrane computing optimization algorithm for the sake of getting the best combination parameters of SVM for classification. This method overcomes the insufficiency of the conventional method which converged to local optimum, at the same time also has the advantages of good robustness, fast convergence speed and obtains the global optimal solution. Finally, to show the applicability and superiority of the proposed algorithm, the method is employed to identify abnormal signal of c-band radio (including radar, jammer, single carrier and single frequency point). Compared with Genetic Algorithm-based SVM (GA-SVM), Simulated Annealing algorithm-based SVM (SA-SVM), Ant Colony algorithm-based SVM (AC-SVM), the proposed model performs best for the four abnormal signal.