This repository provides a bookdown GitHub template for completing the end of chapter "Conceptual" and "Applied" exercises from the the book An Introduction to Statistical Learning by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani.
- Generate a new repository from this template.
- Fill in your answers in R Markdown between the commented questions.
- Push your edits to build your bookdown book of solutions.
To complete your solutions, please generate a new repository from this template this repository first. You can then add solutions between the relevant commented questions using R markdown.
All questions from the original book have been translated to markdown and are quoted. You should provide your answers between the commented sections. For example:
> For each of parts (a) through (d), indicate whether we would generally expect
> the performance of a flexible statistical learning method to be better or
> worse than an inflexible method. Justify your answer.
>
> a. The sample size n is extremely large, and the number of predictors p is
> small.
In this case, a flexible method would be better because we have a large number
of observations so can more reliably detect subtle patterns in the data and are
likely to avoid overfitting due to the low number of predictors.
> b. The number of predictors p is extremely large, and the number of
> observations n is small.
Erm, perhaps the opposite?You may want to edit this README.md (to reflect that your version is not a
template!), and edit the edit tag in _output.yml to point to your
repository.
For some questions, you are required to read in data. These data are provided
in subdirectory data, downloaded from the resources section of
https://www.statlearning.com.
This repository is setup using bookdown and can build a gitbook style bookdown
book. Options are controlled in _bookdown.yml and _output.yml. Notably, I've
set new_session: yes to start a new R session for each chapter of the book,
so R packages required for answers should be reloaded for each chapter
that requires them.
To build the book locally you can run:
bookdown::render_book('index.Rmd', 'bookdown::gitbook')
Or hit the "build" button in your RStudio session.
I have specified a .github/workflows directory to implement a GitHub
workflow to build the book on updates to the repository and then host the
built book from GitHub pages. For example, to see the built version of this
template, see the GitHub pages deployment. When generating a new repository
from this template you may need to enable workflows to get this to run. The
workflow will create a new branch gh-pages containing the built book and will
deploy this book for you.
R package dependencies should be specified in the (fake) package DESCRIPTION
file. These dependencies will then be installed by the workflow so that your
book build works as expected.
I've added some (minimal) styling for the book inside islrv2.css, notably,
using Computer Modern in homage to the the book (and the rest of this font's
history!). To achieve this, I've made use of a web version of computer modern.
Many thanks for the original authors for a great book on statistical learning.
If you notice any errors, feel free to make a pull request.