Solution path algorithm for twin multi-class support vector machine

The twin support vector machine and its extensions have made great achievements in dealing with binary classification problems. However, it suffers from difficulties effective solution of multi-classification fast model selection. This work devotes to the regularization parameter tuning algorithm for multi-class machine. Specifically, a novel sample data set partition strategy is first adopted, which basis construction. Then, combining linear equations block matrix theory, Lagrangian multipliers are proved be piecewise w.r.t. parameters, so that parameters continuously updated by only solving break points. Next, 1 as approaches infinity, thus, simple yet initialization devised. Finally, eight kinds events defined seek starting event next iteration. Extensive experimental results on nine UCI sets show proposed method can achieve comparable performance without any quadratic programming problem.