New mixed integer fractional programming problem and some multi-objective models for sparse optimization

Abstract We propose a novel Mixed-Integer Nonlinear Programming (MINLP) model for sparse optimization based on the polyhedral k-norm. put special emphasis application of in Feature Selection Support Vector Machine (SVM) classification. address continuous relaxation problem, which comes out form fractional programming problem (FPP). In particular, we consider possible way tackling FPP by reformulating it via DC (Difference Convex) decomposition. also overview SVM models and related terms multi-objective optimization. The results some numerical experiments benchmark classification datasets are reported.