hanley {gcl} | R Documentation |
Compares two binary classifier outputs on the same data using the method by Hanley and McNeil.
hanley(paref1, paref2, aaref, ctype = "pearson")
paref1 |
vector of predictions from classifier 1 |
paref2 |
vector of predictions from classifier 2 |
aaref |
actual classes, must be 0 or 1 |
ctype |
type of correlation used in cor.test |
Implements the method of Hanley and McNeil for comparing two c-indices (ROC AUC's) from the same data.
Returns a list.
p |
p value |
auc1 |
the AUC of classifier 1 |
auc2 |
the AUC of classifier 2 |
se1 |
standard error of AUC of classifier 1 |
se2 |
standard error of AUC of classifier 2 |
z |
z value |
r |
Correllation factor |
corr0s |
correlation of true 0's |
corr1s |
correlation of true 1's |
Staal A. Vinterbo staal@dsg.harvard.edu
J.A. Hanley and B.J. McNeil. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology, 148:839-43, 1983.
p1 <- runif(10) p2 <- runif(10) a <- round(runif(10)) hanley(p1,p2,a)