Skip to content

Can't calculate 95% CI for AUC #4

@dhananyashunmugam

Description

@dhananyashunmugam

I have 3 classes , using multiROC i have calculated AUC for class1,2,3, macro AUC and micro AUC. I want to calculate 95% CI for all AUC. My dataset is imbalance so I have used roc_auc_with_ci_strata function to calculate that. And also I have AUC for class1=1, class2=1, class3=1, Micro AUC=1, Macro AUC=0.97. When i try the function, I'm getting this error,

"All values of t are equal to 1 \n Cannot calculate confidence intervals"
Error in x[[jj]][iseq] <- vjj : replacement has length zero

When I check the code,

res_boot

STRATIFIED BOOTSTRAP

Call:
boot(data = final_df, statistic = multi_roc_auc, R = R, strata = final_df$class)

Bootstrap Statistics :
original bias std. error
t1* 1.0000000 0 0
t2* 1.0000000 0 0
t3* 1.0000000 0 0
t4* 0.9770418 0 0
t5* 1.0000000 0 0

Kindly help to solve this error

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions