https://github.com/jtleek/slipper

Tidy and easy bootstrapping
https://github.com/jtleek/slipper

Last synced: 1 day ago
JSON representation

Tidy and easy bootstrapping

• Host: GitHub
• URL: https://github.com/jtleek/slipper
• Owner: jtleek
• Created: 2016-12-08T21:08:11.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-10-05T18:12:29.000Z (about 7 years ago)
• Last Synced: 2024-08-02T06:03:34.118Z (3 months ago)
• Language: R
• Size: 72.3 KB
• Stars: 118
• Watchers: 7
• Forks: 12
• Open Issues: 2
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
Bootstrapping made easy and tidy with slipper

=================

![](slipper.jpg)

You've heard of [broom](https://cran.r-project.org/web/packages/broom/index.html) for tidying up your R functions. slipper is an R package for tidy/easy bootstrapping. There are already a bunch of good bootstrapping packages out there including [bootstrap](https://cran.r-project.org/web/packages/bootstrap/) and [boot](https://cran.r-project.org/web/packages/boot/). You can also bootstrap with [dplyr and broom](https://cran.r-project.org/web/packages/broom/vignettes/bootstrapping.html) or with [purrr and modelr](https://cran.r-project.org/web/packages/modelr).

But I'm too dumb for any of those. So slipper includes some simple,pipeable bootstrapping functions for me 

### install

with `devtools`:

```r

devtools::install_github('jtleek/slipper')

```

### use

There are only two functions in this package. 

Call `slipper` to bootstrap any function that returns

a single value. 

```r

slipper(mtcars,mean(mpg),B=100)

```

slipper is built to work with pipes and the tidyverse too. 

```r

mtcars %>% slipper(mean(mpg),B=100)

```

The output is a data frame with the values of the function on the original data set and the bootstrapped replicates. You can calculate confidence intervals using summarize

```r

mtcars %>% slipper(mean(mpg),B=100) %>%

  filter(type=="bootstrap") %>% 

  summarize(ci_low = quantile(value,0.025),

            ci_high = quantile(value,0.975))

```

You can also bootstrap linear models using `slipper_lm` just pass the data frame and the formula you want to fit on the original data and on the bootstrap samples. 

```r

 slipper_lm(mtcars,mpg ~ cyl,B=100)

```

This is also pipeable

```r

mtcars %>% slipper_lm(mpg ~ cyl,B=100)

```

The default behavior is to bootstrap complete cases, but if you want to bootstrap residuals set `boot_resid=TRUE`

```S

mtcars %>% slipper_lm(mpg ~ cyl,B=100,boot_resid=TRUE)

```

You can calculate bootstrap confidence intervals in the same way as you do for `slipper`.

```r

mtcars %>% slipper_lm(mpg ~ cyl,B=100) %>% 

 filter(type=="bootstrap",term=="cyl") %>%

  summarize(ci_low = quantile(value,0.025),

            ci_high = quantile(value,0.975))

```

Finally if you want to do a bootstrap hypothesis test you can pass a formula and a nested null formula. `formula` must every term in `null_formula` and one additional one you want to test. 

```r

# Bootstrap hypothesis test - 

# here I've added one to the numerator

# and denominator because bootstrap p-values should 

# never be zero.

mtcars %>% 

  slipper_lm(mpg ~ cyl, null_formula = mpg ~ 1,B=1000) %>%

    filter(term=="cyl") %>%

    summarize(num = sum(abs(value) >= abs(value[1])),

                                den = n(),

                                pval = num/den)

```

That's basically it for now. Would love some help/pull requests/fixes as this is my first attempt at getting into the tidyverse :).