compute_wt: Weight value from concave function

In mpath: Regularized Linear Models

Weight value from concave function

Description

Weight value from concave function

Usage

compute_wt(z, weights, cfun, s, delta=0.0001)

Arguments

z	vector nonnegative values from dfun, e.g., u^2/2
weights	optional numeric vector of weights.
cfun	integer from 1-8, concave function as in irglm_fit
s	a numeric value, see details in irglm_fit
delta	a positive small value, see details in irglm_fit

Value

Weight value from concave function

Author(s)

Zhu Wang <[e-mail]>

References

Zhu Wang (2020) Unified Robust Estimation, arXiv e-prints, https://arxiv.org/abs/2010.02848

See Also

irglmreg

Examples

u <- seq(0, 4, by=0.01)
z <- u^2/2 ### this is dfun
res <- compute_wt(z, cfun=1, s=1)
plot(z, res, ylab="Weight", type="l", lwd=2, 
     main=expression(paste("hcave", "   (", sigma, "=1)", )))

mpath documentation built on Jan. 7, 2023, 1:17 a.m.