SLOPE 0.2.0
Introduction to SLOPE
SLOPE (Bogdan et al., 2015) stands for sorted L1 penalized estimation and is a generalization of OSCAR (Bondell & Reich, 2008). As the name suggests, SLOPE is a type of $\ell_1$regularization. More specifically, SLOPE fits generalized linear models regularized with the sorted $\ell_1$ norm. The objective in SLOPE is
\[\operatorname{minimize}\left\{ f(\beta) + J(\beta \mid \lambda)\right\},\]where $f(\beta)$ is typically the loglikelihood of some model in the family of generalized linear models and
\[J(\beta\mid \lambda) = \sum_{i=1}^p \lambda_i\beta_{(i)}\]is the sorted $\ell_1$ norm.
Some people will note that this penalty is a generalization of the standard $\ell_1$ norm penalty^{1}. As such, SLOPE is a type of sparse regression—just like the lasso. Unlike the lasso, however, SLOPE gracefully handles correlated features. Whereas the lasso often discards all but a few among a set of correlated features (Jia & Yu, 2010), SLOPE instead clusters such features together by setting such clusters to have the same coefficient in absolut value.
SLOPE 0.2.0
SLOPE 0.2.0 is a new verison of the R package SLOPE featuring a range of improvements over the previous package. If you are completely new to the package, please start with the introductory vignette.
More model families
Previously, SLOPE only features ordinary leastsquares regression. Now the
package features logistic, Poisson, and multinomial regression on top of that.
Just as in other similar packages, this is enabled simply by
setting family = "binomial"
for logistic regression, for instance.
library(SLOPE)
fit < SLOPE(wine$x, wine$y, family = "multinomial")
Regularization path fitting
By default, SLOPE now fits a full regularization path instead of only a single penalty sequence at once. This behavior is now analogous with the default behavior in glmnet.
plot(fit)
Predictor screening rules
The package now uses predictor screening rules to vastly improve performance in the $p \gg n$ domain. Screening rules are part of what makes other related packages such as glmnet so efficient. In SLOPE, we use a variant of the strong screening rules for the lasso (Tibshirani et al., 2012).
xy < SLOPE:::randomProblem(100, 1000)
system.time({SLOPE(xy$x, xy$y, screen = TRUE)})
## user system elapsed
## 21.161 0.086 3.331
system.time({SLOPE(xy$x, xy$y, screen = FALSE)})
## user system elapsed
## 6.538 0.035 0.941
Crossvalidation and caret
There is now a function trainSLOPE()
, which can be used to run
crossvalidation for optimal selection of sigma
and q
. Here, we run
8fold crossvalidation repeated 5 times.
# 8fold crossvalidation repeated 5 times
tune < trainSLOPE(subset(mtcars, select = c("mpg", "drat", "wt")),
mtcars$hp,
q = c(0.1, 0.2),
number = 8,
repeats = 5)
plot(tune)
In addition, the package now also features a function caretSLOPE()
that
can be used via the excellent caret package, which enables a swath
of resampling methods and comparisons.
C++ and ADMM
All of the performancecritical code for SLOPE has been rewritten in
C++. In addition, the package now features an ADMM solver for
family = "gaussian"
, enabled by setting solver = "admm"
in the call
to SLOPE()
. Preliminary testing shows that this solver is faster for
many designs, particularly when there is high correlation among predictors.
Sparse design matrices
SLOPE now also allows sparse design matrcies of classes from the Matrix package.
And much more…
For a full list of changes, please see the changelog.
References

Simply set $\lambda_i = \lambda_j$ for all $i,j \in {1,\dots,p}$ and you get the lasso penalty. ↩