conf_fdr {bvhar} | R Documentation |
Evaluate the Sparsity Estimation Based on FDR
Description
This function computes false discovery rate (FDR) for sparse element of the true coefficients given threshold.
Usage
conf_fdr(x, y, ...)
## S3 method for class 'summary.bvharsp'
conf_fdr(x, y, truth_thr = 0, ...)
Arguments
x |
|
y |
True inclusion variable. |
... |
not used |
truth_thr |
Threshold value when using non-sparse true coefficient matrix. By default, |
Details
When using this function, the true coefficient matrix should be sparse.
False discovery rate (FDR) is computed by
where TP is true positive, and FP is false positive.
Value
FDR value in confusion table
References
Bai, R., & Ghosh, M. (2018). High-dimensional multivariate posterior consistency under global–local shrinkage priors. Journal of Multivariate Analysis, 167, 157–170.
See Also
[Package bvhar version 2.0.1 Index]