loss3 {mpath} | R Documentation |
Composite Loss Value for GLM
Description
Compute composite loss value
Usage
loss3(y, mu, theta, weights, cfun, family, s, delta)
Arguments
y |
response variable values, 0/1 if |
mu |
response prediction of |
theta |
scale parameter for |
weights |
observation weights, same length as |
cfun |
integer from 1-8, concave function as in |
family |
integer 2, 3 or 4, convex function binomial, Poisson or negative binomial, respectively |
s |
tuning parameter of |
delta |
a small positive number provided by user only if |
Details
For large s
values, the loss can be 0 with cfun=2,3,4
, or "acave", "bcave", "ccave".
Value
Weighted loss values
Author(s)
Zhu Wang <zwang145@uthsc.edu>
References
Zhu Wang (2024) Unified Robust Estimation, Australian & New Zealand Journal of Statistics. 66(1):77-102.
See Also
loss2
irglm
irglmreg
loss2_irsvm