不均衡数据分类器分类性能AUC与Accuracy的比较  

Comparison of the Classification Performance AUC and Accuracy of Classifiers Based on Unbalanced Data

作　　者：赵存秀[1] ZHAO Cun-xiu(Department of Computing,Technology and Business College,Taiyuan 030006,China)

机构地区：[1]山西工商学院计算机信息工程学院,山西太原030006

出　　处：《唐山师范学院学报》2019年第6期75-77,132,共4页Journal of Tangshan Normal University

摘　　要：针对不均衡数据,借助已有的评价指标一致性(consistent)和区分度(discriminating),比较Logistic和LDA学习算法的评价方法AUC和精确率,结果表明,AUC用于学习算法的估计比精度率好。To solve the binary classification problem,accuracy is generally used to evaluate the classification performance of classifiers.In recent years,the area under the ROC(Receiver Operating Characteristics)curve,or simply AUC,has been used to evaluate the classifier classification performance.It avoids the supposed subjectivity in the threshold selection process and provides a single-number“summary”for the performance of the learning algorithms.It has been proved AUC a better measure than accuracy in balance data.In the paper,using consistent and discriminating proposed to comparing the AUC and accuracy of Logistic and LDA.Then we present the empirical estimation,and we get that the estimation of AUC for learning algorithm is better than the precision rate from the experiment.

关 键 词：LOGISTIC LDA学习算法 不均衡 AUC 精确率 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]