Support Vector Machines for uncertainty region detection

A new technique for the detection of the uncertainty region in classification problems is presented. The core of the method is the determination of the best supporting hyperplane for convex hulls of sets of points in a multidimensional input space. To this aim a modified version of the algorithm for the Generalized Optimal Hyperplane is shown to be effective. As in the Support Vector Machine approach, kernel functions can be used to generalize the proposed technique, so as to detect uncertainty regions with nonlinear boundaries. Simulations concerning artificial benchmarks show the quality of the resulting method.