Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/tlverse/sl3

💪 🤔 Modern Super Learning with Machine Learning Pipelines
https://github.com/tlverse/sl3

data-science ensemble-learning ensemble-model machine-learning model-selection r r-package regression stacking statistics

Last synced: about 2 months ago
JSON representation

💪 🤔 Modern Super Learning with Machine Learning Pipelines

Awesome Lists containing this project

README 

        
---

always_allow_html: true

output:

  rmarkdown::github_document

bibliography: "inst/REFERENCES.bib"

---



```{r, echo = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "README-"

)

```



# R/`sl3`: Super Machine Learning with Pipelines



[![R-CMD-check](https://github.com/tlverse/sl3/workflows/R-CMD-check/badge.svg)](https://github.com/tlverse/sl3/actions)

[![Coverage Status](https://codecov.io/gh/tlverse/sl3/branch/master/graph/badge.svg)](https://codecov.io/gh/tlverse/sl3)

[![Project Status: Active – The project has reached a stable, usable state and is being actively developed.](https://www.repostatus.org/badges/latest/active.svg)](https://www.repostatus.org/#active)

[![License: GPL v3](https://img.shields.io/badge/License-GPL%20v3-blue.svg)](https://www.gnu.org/licenses/gpl-3.0)

[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.1342293.svg)](https://doi.org/10.5281/zenodo.1342293)



> A flexible implementation of the Super Learner ensemble machine learning

> system



__Authors:__ [Jeremy Coyle](https://github.com/jeremyrcoyle),

[Nima Hejazi](https://nimahejazi.org),

[Ivana Malenica](https://github.com/imalenica),

[Rachael Phillips](https://github.com/rachaelvp), and

[Oleg Sofrygin](https://github.com/osofr)



---



## What's `sl3`?



`sl3` is an implementation of the Super Learner ensemble machine learning

algorithm of @vdl2007super. The Super Learner algorithm performs ensemble

learning in one of two fashions:



1. The _discrete_ Super Learner can be used to select the best prediction

   algorithm from among a supplied library of machine learning algorithms

   ("learners" in the `sl3` nomenclature) -- that is, the discrete Super Learner

   is the single learning algorithm that minimizes the cross-validated risk.

2. The _ensemble_ Super Learner can be used to assign weights to a set of

   specified learning algorithms (from a user-supplied library of such

   algorithms) so as to create a combination of these learners that minimizes

   the cross-validated risk. This notion of weighted combinations has also been 

   referred to as _stacked regression_ [@breiman1996stacked] and 

   _stacked generalization_ [@wolpert1992stacked].



Looking for long-form documentation or a walkthrough of the `sl3` package?

Don't worry! Just browse [the chapter in our

book](https://tlverse.org/tlverse-handbook/sl3.html).



---



## Installation



Install the _most recent version_ from the `master` branch on GitHub via

[`remotes`](https://CRAN.R-project.org/package=remotes):



```{r gh-master-installation, eval = FALSE}

remotes::install_github("tlverse/sl3")

```



Past stable releases may be located via the

[releases](https://github.com/tlverse/sl3/releases) page on GitHub and may be

installed by including the appropriate major version tag. For example,



```{r gh-version-installation, eval = FALSE}

remotes::install_github("tlverse/[email protected]")

```



To contribute, check out the `devel` branch and consider submitting a pull

request.



---



## Issues



If you encounter any bugs or have any specific feature requests, please [file an

issue](https://github.com/tlverse/sl3/issues).



---



## Examples



`sl3` makes the process of applying screening algorithms, learning algorithms,

combining both types of algorithms into a stacked regression model, and

cross-validating this whole process essentially trivial. The best way to

understand this is to see the `sl3` package in action:



```{r sl3-simple-example, message=FALSE, warning=FALSE}

set.seed(49753)

library(tidyverse)

library(data.table)

library(SuperLearner)

library(origami)

library(sl3)



# load example data set

data(cpp)

cpp <- cpp %>%

  dplyr::filter(!is.na(haz)) %>%

  mutate_all(~ replace(., is.na(.), 0))



# use covariates of intest and the outcome to build a task object

covars <- c("apgar1", "apgar5", "parity", "gagebrth", "mage", "meducyrs",

            "sexn")

task <- sl3_Task$new(

  data = cpp,

  covariates = covars,

  outcome = "haz"

)



# set up screeners and learners via built-in functions and pipelines

slscreener <- Lrnr_pkg_SuperLearner_screener$new("screen.glmnet")

glm_learner <- Lrnr_glm$new()

screen_and_glm <- Pipeline$new(slscreener, glm_learner)

SL.glmnet_learner <- Lrnr_pkg_SuperLearner$new(SL_wrapper = "SL.glmnet")



# stack learners into a model (including screeners and pipelines)

learner_stack <- Stack$new(SL.glmnet_learner, glm_learner, screen_and_glm)

stack_fit <- learner_stack$train(task)

preds <- stack_fit$predict()

head(preds)

```



### Parallelization with `future`s



While it's straightforward to fit a stack of learners (as above), it's easy to

take advantage of `sl3`'s built-in parallelization support too. To do this,

you can simply choose a `plan()` from the [`future`

ecosystem](https://CRAN.R-project.org/package=future).



```{r sl3-parallel-example, eval=FALSE, message=FALSE, warning=FALSE}

# let's load the future package and set 4 cores for parallelization

library(future)

plan(multicore, workers = 4L)



# now, let's re-train our Stack in parallel

stack_fit <- learner_stack$train(task)

preds <- stack_fit$predict()

```



### Controlling the number of CV folds



In the above examples, we fit stacks of learners, but didn't create a Super

Learner ensemble, which uses cross-validation (CV) to build the ensemble model.

For the sake of computational expedience, we may be interested in lowering the

number of CV folds (from 10).  Let's take a look at how to do both below.



```{r sl3-folds-example, eval=FALSE, message=FALSE, warning=FALSE}

# first, let's instantiate some more learners and create a Super Learner

mean_learner <- Lrnr_mean$new()

rf_learner <- Lrnr_ranger$new()

sl <- Lrnr_sl$new(mean_learner, glm_learner, rf_learner)



# CV folds are controlled in the sl3_Task object; we can lower the number of

# folds simply by specifying this in creating the Task

task <- sl3_Task$new(

  data = cpp,

  covariates = covars,

  outcome = "haz",

  folds = 5L

)



# now, let's fit the Super Learner with just 5-fold CV, then get predictions

sl_fit <- sl$train(task)

sl_preds <- sl_fit$predict()

```



The `folds` argument to `sl3_Task` supports both integers (for V-fold CV) and

all of the CV schemes supported in the [`origami`

package](https://CRAN.R-project.org/package=origami). To see the full list,

query `?fold_funs` from within `R` or take a look at [`origami`'s online

documentation](https://tlverse.org/origami/reference/).



---



## Learner Properties



Properties supported by `sl3` learners are presented in the following table:



```{r sl3-learner-properties, warning=FALSE, echo=FALSE, message=FALSE}

library(sl3)

library(kableExtra)



all_properties <- sl3_list_properties()



get_all_learners <- function() {

  # search for objects named like sl3 learners

  learner_names <- apropos("^Lrnr_")

  learners <- mget(learner_names, inherits = TRUE)



  # verify that learner inherits from Lrnr_base (and is therefore an actual

  # sl3 Learner)

  is_learner_real <- sapply(learners, `[[`, "inherit") == "Lrnr_base"

  return(learners[which(is_learner_real)])

}

all_learners <- get_all_learners()



# get a list where each element is a list of properties of a particular learner

get_learner_class_properties <- function(learner_class) {

  return(learner_class$private_fields$.properties)

}

learner_property_lst <- lapply(all_learners, get_learner_class_properties)



# get a list where each element is a list of logicals indicating whether a

# learner has a property or not

get_learner_class_properties_ownership <- function(properties) {

  return(sl3_list_properties() %in% unlist(properties))

}

learner_property_ownership <- lapply(learner_property_lst,

                                     get_learner_class_properties_ownership)



get_learner_name <- function(learner_class) {

  return(learner_class$classname)

}

learner_names <- lapply(all_learners, get_learner_name)



# generate matrix whose columns are properties and rows are learner names

final_mat <- matrix(unlist(learner_property_ownership),

                    nrow = length(learner_property_ownership),

                    ncol = length(sl3_list_properties()), byrow = TRUE)

dimnames(final_mat) <- list(unlist(learner_names), sl3_list_properties())

final_mat.df <- as.data.frame(final_mat)



char_final_mat.df <- ifelse(final_mat.df == TRUE, "√", "x")

kable(char_final_mat.df) %>%

  kable_styling(bootstrap_options = c("striped", "hover", "condensed",

                                      "responsive")) %>%

  scroll_box(width = "100%", height = "200px", fixed_thead = T)

```



---



## Contributions



Contributions are very welcome. Interested contributors should consult our

[contribution

guidelines](https://github.com/tlverse/sl3/blob/master/CONTRIBUTING.md) prior to

submitting a pull request.



---



## Citation



After using the `sl3` R package, please cite the following:



        @software{coyle2021sl3-rpkg,

          author = {Coyle, Jeremy R and Hejazi, Nima S and Malenica, Ivana and

            Phillips, Rachael V and Sofrygin, Oleg},

          title = {{sl3}: Modern Pipelines for Machine Learning and {Super

            Learning}},

          year = {2021},

          howpublished = {\url{https://github.com/tlverse/sl3}},

          note = {{R} package version 1.4.2},

          url = {https://doi.org/10.5281/zenodo.1342293},

          doi = {10.5281/zenodo.1342293}

        }



---



## License



© 2017-2021 [Jeremy R. Coyle](https://github.com/jeremyrcoyle), [Nima S.

Hejazi](https://nimahejazi.org), [Ivana

Malenica](https://github.com/podTockom), [Rachael V.

Phillips](https://github.com/rachaelvp), [Oleg

Sofrygin](https://github.com/osofr)



The contents of this repository are distributed under the GPL-3 license. See

file `LICENSE` for details.



---



## References