auc {enrichvs} | R Documentation |
Function to calculate the Area Under the ROC Curve (AUC)
auc(x, y, decreasing=TRUE, top=1.0)
x |
a vector for scores |
y |
a vector for labels |
decreasing |
TRUE if the compounds are ranked by decreasing score |
top |
threshold ratio of the false positives (when ROC analysis is performed on a top list) |
AUC, in the range from 0 to 1.
Hiroaki YABUUCHI
Truchon et al. Evaluating Virtual Screening Methods: Good and Bad Metrics for the "Early Recognition" Problem. J. Chem. Inf. Model. (2007) 47, 488-508.
x <- rnorm(1000) # random scores for 1000 compounds y <- c(rep(1,50), rep(0,950)) # activity labels for "x" auc(x, y) data(dud_egfr) auc(dud_egfr$energy, dud_egfr$label, decreasing=FALSE) # AUC up to the first 10% of the false positives (ROC10%) auc(dud_egfr$energy, dud_egfr$label, decreasing=FALSE, top=0.1)