A new multi-class classification method based on minimum enclosing balls

摘要: With respect to classification problems, the Minimum enclosing ball (MEB) method was recently studied by some scholars as a new support vector machine. As nascent technology, however, MEB reports poor adaptability for different types of samples, especially multi-class samples. In this paper, we propose based on MEB. This is derived from each class sample center and radius with Gaussian kernel width factor parameter σ, which labelled σ-MEB. σ variable according characteristics. When considered, classifier easy adapt robust in diverse datasets. The quadratic programming problem transformed into its dual form Lagrange multipliers using method. Finally, applied sequential minimal optimization Karush—Kuhn—Tucker conditions accelerate training process. Numerical experiment results indicate that given proposed more accurate than methods it compared. Moreover, values upper quantile adaptation performance.