Support vector machines for multi-class classification

摘要: Support vector machines (SVMs) are primarily designed for 2-class classification problems. Although in several papers it is mentioned that the combination of K SVMs can be used to solve a K-class problem, such procedure requires some care. In this paper, scaling problem different highlighted. Various normalization methods proposed cope with and their efficiencies measured empirically. This simple way ssing learn consists choosing maximum applied outputs solving one-per-class decomposition general problem. second part more sophisticated techniques suggested. On one hand, stacking other proposed. end, scheme replaced by elaborated schemes based on error-correcting codes. An incremental algorithm elaboration pertinent mentioned, which exploits properties an efficient computation.