In this algorithm the classification of samples in the critical zone is decided by the voting of K-NN set in the feature space and the sample space.

• 该方法将特征空间分类超平面附近的样本分类，交由特征空间和样本空间中的K-近邻集体投票表决。