A Novel Multiclass Classifier Based on the Analytical Center of Version Space
Zhengding Qiu · Fudan xuebao. Ziran Kexue ban · 2004
Considering that for the one versus all(OvA) approach, repeating construction of all classifier leads to daunting computation and low efficiency of classification, and multi-class classifier based on SVM is not very effective when the version space is asymmetric or elongated. Those problems are addressed by proposing a multi-class classifier based on the analytical center of version space, which is called MACM. Experiments on wine recognition and glass identification dataset demonstrate that MACM is validated.