Optimizing the classification cost using SVMs with a double hinge loss

URN:NBN:SI:DOC-59V9T4902014Ahmed, Amiroudocuments/doc/5/URN_NBN_SI_doc-59V9T490_001.pdfdocuments/doc/5/URN_NBN_SI_doc-59V9T490_001.txt238articlestr. 125-1330350-559628507431URN:NBN:SI:doc-59V9T490engSlovensko društvo InformatikaInformatica (Ljubljana)InCmetoda podpornih vektorjevoptimizacijastrojno učenjeOptimizing the classification cost using SVMs with a double hinge loss