Spectral Signature Classification Using A Support Vector Classifier For Real-Time Instrumentation

The research WSR-88D (weather surveillance radar) locally operated by the National Severe Storm Laboratory (NSSL) in Norman has the unique capability of collecting massive volumes of Level I time series data over many hours which provides a rich environment for evaluating our new post-processing algorithms. In this work, a Support Vector Machine (SVM) classifier is employed to identify tornado vortices based on their characteristic Doppler spectra and eigen analysis technique. A SVM-based classifier evades the pitfalls of the traditional statistical learning algorithms, such as neural networks, by setting up a convex optimization problem with a single global minimum. In addition, through the use of kernels and nonlinear mapping to higher dimensional spaces, the SVM classifier is able to effectively handle nonlinear classification problems. Finally, the SVM classifier has the added advantage of reducing overfitting by constructing a maximum margin separating hyperplane in a higher dimensional feature space which ensures a small generalization error bound .