Decision-level Fusion Scheme of SVM and naive Bayes Classifier for Radar Target Recognition

In radar target recognition, both the noise robustness and the recognition accuracy are important problems. We proposed the new scheme obtaining improved result by combining two classifiers. One of classifiers is robust to noise, another is more accurate. Noise robustness and accuracy are improved by selecting the suitable classifier at the decision level. Simulation result shows that the recognition performance and noise robustness of the proposed classifier are improved compared to those of individual classifier when the combining technique is used.