Marginal false discovery rates

Estimates the marginal false discovery rate (mFDR) of a group penalized regression model.

Usage

mfdr(fit, X)

Arguments

fit

A grpreg or grpsurv object.

X

The model matrix corresponding to fit. This is not necessary for linear regression, but in logistic and Cox regression, the mFDR depends on X. It is not necessary to supply X if it is already contained in fit; i.e., if ncvreg/ncvsurv was run with returnX = TRUE.

Value

An object with S3 class mfdr inheriting from data.frame, containing:

ef

The number of variables selected at each value of lambda, averaged over the permutation fits.

s

The actual number of selected variables for the non-permuted data.

mfdr

The estimated marginal false discovery rate (ef/s).

Details

The function estimates the marginal false discovery rate (mFDR) for groups in a group lasso or group MCP penalized regression model. The estimate tends to be accurate in most settings, but will be somewhat conservative if predictors are highly correlated.

See also

grpreg(), grpsurv()

Examples

# Birthweight data ---------------------------
data(Birthwt)
x <- Birthwt$X
group <- Birthwt$group

# Linear regression --------------------------
y <- Birthwt$bwt
fit <- grpreg(x, y, group)
obj <- mfdr(fit)
head(obj)
#>                   ef s        mfdr
#> 0.206495 0.000000000 0 0.000000000
#> 0.188151 0.001073303 1 0.001073303
#> 0.171436 0.003251466 1 0.003251466
#> 0.156206 0.008496796 1 0.008496796
#> 0.142329 0.019651242 1 0.019651242
#> 0.129685 0.040878376 2 0.020439188

# Logistic regression ------------------------------
y <- Birthwt$low
fit <- grpreg(x, y, group, penalty="grMCP", family="binomial", returnX = TRUE)
obj <- mfdr(fit)
# If returnX is not TRUE, user must supply X
fit <- grpreg(x, y, group, penalty="grMCP", family="binomial")
obj <- mfdr(fit, x)
head(obj)
#>                   ef s       mfdr
#> 0.0960555 0.00000000 0 0.00000000
#> 0.0875222 0.03048445 1 0.03048445
#> 0.0797470 0.06129080 1 0.06129080
#> 0.0726625 0.11468468 1 0.11468468
#> 0.0662073 0.20446253 3 0.06815418
#> 0.0603256 0.34473624 4 0.08618406

# Cox regression -----------------------------------
data(Lung)
x_lung <- Lung$X
y_lung <- Lung$y
g_lung <- Lung$group
fit <- grpsurv(x_lung, y_lung, g_lung, penalty = "grSCAD")
obj <- mfdr(fit, x_lung)
head(obj)
#>                 ef s        mfdr
#> 0.2632 0.000000000 0 0.000000000
#> 0.2455 0.005998595 1 0.005998595
#> 0.2290 0.012415206 1 0.012415206
#> 0.2135 0.024215529 1 0.024215529
#> 0.1991 0.044357731 2 0.022178865
#> 0.1857 0.077020234 2 0.038510117