-
``` r
library(nlme) ## mixed models with more advanced covariance structures
library(lme4) ## mixed models for glm
library(mlogit) ## multinomial (mixed) logistic regression
library(mnlogit) ## multi…
-
It would be very useful to be able to perform variable selection with competing risks. It would seem that it should be straightforward, since `glmnet` has a `multinomial` family. But as we've discover…
-
search in our docs for screening only shows the release announcement as related topic
gist from PR #4683 looks okay
https://gist.github.com/josef-pkt/b94cbeb1a0663930bf0aa5381b3a1064
stackoverf…
-
shrinking the endog is another principle that allows reuse of existing methods for robust regression. This is similar to winsorizing and an alternative to trimming or dropping outliers (e.g. #3273 #9…
-
### The issue
The `PenalizedLeastSquaresAlgorithm` class can manage a penalized least squares problem, taking into account for a non zero penalization factor:
https://github.com/openturns/opentu…
-
Sparse Laplacian Shrinkage combines a L1 based penalty and a quadratic informative penalty, similar to glm-net but with structured L2 penalization matrix
Sparse Laplacian Shrinkage is the first stran…
-
my guess is that this might be similar to regression trees, but I never looked at those details.
The basic idea: We want to select variables from a large possible list. Explanatory variables are gene…
-
R mgcv has a `by` option for the interaction effect of penalized splines and a categorical variables.
I don't know yet how we will support this. patsy can create the interaction term using its spli…
-
subtopic of #4038
(I don't have an overview where the different pieces use for penalization are.)
specifically: Where is a quadratic penalty function, i.e. normal prior?
**Penalization classe…
-
a companion issue to #4257
topic: towards generic local influence functions and measures, more theory and applications
I still don't have a good overview how the different special cases connect.
…