SVM: Related formulations and algorithms

00:28 Hinge primal SVM
6:18 Loss and penalty framework (binomial deviance, SVM hinge, Huberized)
12:45 Another formulation in ESL textbook
17:04 \nu-SVC
20:30 Sparse SVC
22:40 SVM for multi-class classification (OVO, OVA)
27:45 Algorithms (SMO)

Credit: Some figures in the handout are taken from ISLR and ESL textbooks,

Instructor: Jitkomut Songsiri, Chulalongkorn University, Thailand