| doAUCcluster {mitoticFigureCounts} | R Documentation |
Do AUC analysis of clustered data.
This function is based on the analysis described in: Obuchowski NA. "Nonparametric analysis of clustered ROC curve data." Biometrics. 1997: 567-578.
This function is an adaptation of a function downloaded from the Cleveland Clinic Lerner Research Institute Department of Quantitative Health Sciences Software web page.
FILE: https://www.lerner.ccf.org/qhs/software/lib/funcs_clusteredROC.R
WEBPAGE: https://www.lerner.ccf.org/qhs/software/roc_analysis.php
doAUCcluster(predictor1, predictor2 = NULL, response, clusterID, alpha = 0.05, level = NULL, print.all = F)
predictor1 |
a vector containing the predictor for ROC curve 1 |
predictor2 |
a vector containing the predictor for ROC curve 2 |
response |
a vector containing the response for both ROC curves |
clusterID |
a vector containing IDs for the clusters |
alpha |
the type I error rate |
level |
can be used to specify the response level considered positive (if omitted, the second level of the response is selected) |
print.all |
if TRUE, intermediate estimates are printed |
iMRMC users shared the links during a discussion with questions about how to analyze MRMC data that was clustered. https://github.com/DIDSR/iMRMC/issues/147 There is a short pdf tutorial a https://www.lerner.ccf.org/qhs/software/lib/clusteredROC_help.pdf. It exists in the inst/extra/docs folder of the repository. It exists in the extra/docs folder of the installed package.
[list] auc, auc.se, ci.for.auc