来源 包：stats 函数：prop.test官方文档：https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/prop.test 包：epiR 函数：epi.tests官方文档：https://www.rdocumentation.org/packages/epiR/versions/1.0-10/topics/epi.tests 目的 判断两个模型的敏感度、特异度、准确率是否有显著差异 代码 ### 统计学检验 #### TP1：模型一的真正例数量# TN1：模型一的真反例数量# FP1：模型一的假正例数量# FN1：模型一的假反例数量# TP2：模型二的真正例数量# TN2：模型二的真反例数量# FP2：模型二的假正例数量# FN2：模型二的假反例数量x <- c(TN1,TN2); n <- c(TN1+FP1, TN2+FP2) # 特异度prop.test(x,n, correct = T)$p.valuex <- c(TP1,TP2); n <- c(TP1+FN1, TP2+FN2) # 敏感度prop.test(x,n, correct = T)$p.valuex <- c(TP1+TN1,TP2+TN2); n <- c(TP1+TN1+FP1+FN1, TP2+TN2+FP2+FN2) # 准确率prop.test(x,n, correct = T)$p.value### 置信区间求解 ###dat <- as.table(matrix(c(TN1, FN1, FP1, TN1), nrow = 2, byrow = TRUE))epi.tests(dat) 置信区间简易求解方法（感谢课题组机智的小猫咪提供的链接）
http://vassarstats.net/clin1.html?tdsourcetag=s_pctim_aiomsg
参考资料：
https://journals.plos.org/plosmedicine/article/file?type=supplementary&id=info:doi/10.1371/journal.pmed.1002699.s001https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1002699