基于马氏抽样的SVM非平衡数据分类算法的泛化性能研究

本文将样本为独立同分布的情形减弱为一致遍历马氏链的情形去研究了非平衡数据分类算法的泛化性能，提出了基于马氏抽样的SVM非平衡数据分类算法、基于马氏抽样的EDSVM非平衡数据分类算法和基于马氏抽样的SVM-WKNN非平衡数据分类算法.并用UCI数据库中的10个实际不平衡数据集进行数值实验，实验结果表明基于马氏抽样的上述三种算法的错分率均比基于随机抽样的对应算法的错分率要低，且上述三种算法中，基于马氏抽样的SVM-WKNN非平衡数据分类算法的泛化性能最好.

Research on the Generalization Performance of SVM Imbalanced Data Classification Algorithm Based on Markov Sampling

This paper changes the assumption that samples are independent and identically distributed to that samples are uniformly ergodic Markov chains,which make it convenient for us to study the generalization performance of the imbalanced data classification algorithm,and SVM imbalanced data classification algorithm based on Markov sampling,EDSVM imbalanced data classification algorithm based on Markov sampling and SVM-WKNN imbalanced data classification algorithm based on Markov sampling are proposed.The numerical experiments of ten actual imbalanced datasets in the UCI database show that the misclassification rate of the above algorithm based on Markov sampling is lower than that of the corresponding algorithm based on random sampling,and the above three algorithms,SVM-WKNN imbalanced data classification algorithm based on Markov sampling has the best generalization performance.