A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment ...A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment effect and interim result.For hypotheses and reversed hypotheses under normal models,we obtain analytical expressions of the ROC curves of the CCP,find optimal ROC curves of the CCP,investigate the superiority of the ROC curves of the CCP,calculate critical values of the False Positive Rate(FPR),True Positive Rate(TPR),and cutoff of the optimal CCP,and give go/no go decisions at the interim of the optimal CCP.In addition,extensive numerical experiments are carried out to exemplify our theoretical results.Finally,a real data example is performed to illustrate the go/no go decisions of the optimal CCP.展开更多
接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boos...接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boosting改进算法,通过在原始Boosting迭代中引入数据重平衡操作,实现弱学习算法优化目标从精度向AUC的迁移.实验结果表明,较之原始Boosting算法,新算法在AUC度量下能获得更好性能.展开更多
针对借贷过程中的信息不对称问题,为更有效地整合不同的数据源和贷款违约预测模型,提出一种集成学习的训练方法,使用AUC(Area Under Curve)值和Q统计值对学习器的准确性和多样性进行度量,并实现了基于AUC和Q统计值的集成学习训练算法(TA...针对借贷过程中的信息不对称问题,为更有效地整合不同的数据源和贷款违约预测模型,提出一种集成学习的训练方法,使用AUC(Area Under Curve)值和Q统计值对学习器的准确性和多样性进行度量,并实现了基于AUC和Q统计值的集成学习训练算法(TABAQ)。基于个人对个(P2P)贷款数据进行实证分析,发现集成学习的效果与基学习器的准确性和多样性关系密切,而与所集成的基学习器数量相关性较低,并且各种集成学习方法中统计集成表现最好。实验还发现,通过融合借款人端和投资人端的信息,可以有效地降低贷款违约预测中的信息不对称性。TABAQ能有效发挥数据源融合和学习器集成两方面的优势,在保持预测准确性稳步提升的同时,预测的一类错误数量更是进一步下降了4.85%。展开更多
基金supported by the National Social Science Fund of China(Grand No.21XTJ001).
文摘A Receiver Operating Characteristic(ROC)analysis of a power is important and useful in clinical trials.A Classical Conditional Power(CCP)is a probability of a classical rejection region given values of true treatment effect and interim result.For hypotheses and reversed hypotheses under normal models,we obtain analytical expressions of the ROC curves of the CCP,find optimal ROC curves of the CCP,investigate the superiority of the ROC curves of the CCP,calculate critical values of the False Positive Rate(FPR),True Positive Rate(TPR),and cutoff of the optimal CCP,and give go/no go decisions at the interim of the optimal CCP.In addition,extensive numerical experiments are carried out to exemplify our theoretical results.Finally,a real data example is performed to illustrate the go/no go decisions of the optimal CCP.
文摘接收者操作特性(Receiver operating characteristics,ROC)曲线下面积(Area under the ROC curve,AUC)常被用于度量分类器在整个类先验分布上的总体分类性能.原始Boosting算法优化分类精度,但在AUC度量下并非最优.提出了一种AUC优化Boosting改进算法,通过在原始Boosting迭代中引入数据重平衡操作,实现弱学习算法优化目标从精度向AUC的迁移.实验结果表明,较之原始Boosting算法,新算法在AUC度量下能获得更好性能.
文摘针对借贷过程中的信息不对称问题,为更有效地整合不同的数据源和贷款违约预测模型,提出一种集成学习的训练方法,使用AUC(Area Under Curve)值和Q统计值对学习器的准确性和多样性进行度量,并实现了基于AUC和Q统计值的集成学习训练算法(TABAQ)。基于个人对个(P2P)贷款数据进行实证分析,发现集成学习的效果与基学习器的准确性和多样性关系密切,而与所集成的基学习器数量相关性较低,并且各种集成学习方法中统计集成表现最好。实验还发现,通过融合借款人端和投资人端的信息,可以有效地降低贷款违约预测中的信息不对称性。TABAQ能有效发挥数据源融合和学习器集成两方面的优势,在保持预测准确性稳步提升的同时,预测的一类错误数量更是进一步下降了4.85%。
