Aggregated Knockoff Using Selection Frequency

Performs Aggregated Knockoff analysis by selecting variables based on their selection frequency across multiple knockoff filters.

agg_Freq(Ws_mat, fdr = 0.05, offset = 0, tau = 0.5)

Arguments

Ws_mat

A matrix where each row represents the test statistics from a different knockoff filter, and each column corresponds to a variable.

fdr

A numeric value specifying the target false discovery rate (FDR) level. Default is \(0.05\).

offset

An integer (0 or 1) specifying the offset in the threshold calculation. Default is \(0\) (liberal control).

tau

A numeric value indicating the selection frequency threshold for variable inclusion. Default is \(0.5\).

Value

A vector shat containing the indices of selected variables after applying selection frequency and the knockoff filter.

References

Ren, Z., Wei, Y., & Candès, E. (2023). Derandomizing knockoffs. Journal of the American Statistical Association, 118(542), 948-958.

See also

Other aggregate: agg_Avg(), agg_BH()

Examples

set.seed(2024)
p = 100; n = 80
X = generate_X(n=80,p=100)
y <- generate_y(X, p_nn=10, a=3)
Xk = create.shrink_Gaussian(X = X, n_ko = 10)
res1 = knockoff.filter(X, y, Xk, statistic = stat.glmnet_coefdiff,
                       aggregate = agg_BH,
                       offset = 1, fdr = 0.1)
res1
#> Call:
#> knockoff.filter(X = X, y = y, Xk = Xk, statistic = stat.glmnet_coefdiff, 
#>     aggregate = agg_BH, fdr = 0.1, offset = 1)
#> 
#> Selected variables:
#> integer(0)
#> 
#> Frequency of selected variables from 10 knockoff copys:
#>   [1] 10 10 10 10 10  8 10 10  9 10  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
#>  [26]  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
#>  [51]  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
#>  [76]  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
agg_Freq(res1$Ws)
#>  [1]  1  2  3  4  5  6  7  8  9 10