1
$\begingroup$

I have a question about how I can calculate true and false positive rate in a simulation study?

I have seen some articles and thesis by different definitions. One of them is the following one:

TPR=the ratio of the truly selected covariates relative to the number of active covariates

and

FPR= the ratio of the falsely selected covariates relative to the number of inactive covariates.

and other definition:

True positives (TP):
Non-null variables correctly identified as non-null variables.
 False positives (FP):
Null variables incorrectly labeled as non-null variables.
 True negatives (TN):
Null variables correctly identified as null variables.
 False negatives (FN):
Non-null variables incorrectly labeled as null variables.
TPR=TP/(TP+FN)

Can anyone tell me which formula is used in variable selection in simulation studies? And what are the definitions for active and inactive covariates?

Improve this question
$\endgroup$
2
  • $\begingroup$ They both say the same. $\endgroup$
    – user88
    Commented Mar 19, 2013 at 19:40
  • $\begingroup$ Hi @mbq..thanks for your response..and do you know what are the definitions for active and inactive covariates? are the following line correct? active covariates in a simulation study are the non null variables used in the simulation model and inactive covariates are the rest of the variables in the simnulation study which are null and have not used in the simulation model? thanks $\endgroup$
    – user22
    Commented Mar 19, 2013 at 23:50

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.