Skip to content

Latest commit

 

History

History
92 lines (60 loc) · 6.06 KB

README.md

File metadata and controls

92 lines (60 loc) · 6.06 KB

isle: Introduction to Statistical Learning in Economics

License: MIT

isle is an open-source educational resource designed to guide you through the fascinating world of statistical learning, with a focus on its applications in economics. This repository provides a comprehensive set of interactive Jupyter notebooks that implement the concepts and methods presented in the renowned book:

Specifically, isle leverages the Python edition of ISL (ISLP) and its associated labs, adapting them for an economics-focused audience. You'll discover how to apply powerful statistical learning techniques to analyze economic data, build predictive models, and gain valuable insights into economic phenomena.

Why isle?

  • Economics Focus: While based on ISLP, isle provides additional context, examples, and exercises tailored to the field of economics.
  • Interactive Learning: The Jupyter notebooks allow you to experiment with code, modify parameters, and visualize results directly in your browser, making the learning process engaging and intuitive.
  • Based on a Classic: isle builds upon the solid foundation of "An Introduction to Statistical Learning," a widely acclaimed textbook known for its clear explanations and practical approach.
  • Python Implementation: Leverage the power and versatility of Python for statistical learning, using libraries like statsmodels, scikit-learn, pandas, and numpy.
  • In-Browser Execution: Run the notebooks directly in your browser - no complex setup required!
  • Open Source & Community-Driven: isle is open-source, welcoming contributions and collaborations from the economics and data science communities.

What You'll Learn

isle covers a wide range of statistical learning topics, including:

  • Introduction to Statistical Learning: What is statistical learning and how is it applied in economics?
  • Linear Regression: Estimate relationships between economic variables and make predictions.
  • Classification: Build models to classify economic outcomes (e.g., recession/expansion, credit default/no default).
  • Resampling Methods: Assess the accuracy and variability of your models using techniques like cross-validation and bootstrapping.
  • Linear Model Selection and Regularization: Improve model performance and interpretability using techniques like Ridge, Lasso, and Elastic Net.
  • Moving Beyond Linearity: Explore non-linear relationships using polynomial regression, splines, and generalized additive models.
  • Tree-Based Methods: Understand and apply decision trees, random forests, and gradient boosting for economic modeling.
  • Support Vector Machines: Learn how to use SVMs for classification and regression tasks.
  • Deep Learning: Get an introduction to neural networks and their potential applications in economics.
  • Survival Analysis: Analyze time-to-event data, such as the duration of unemployment spells.
  • Unsupervised Learning: Discover patterns and structure in economic data using techniques like principal component analysis and clustering.
  • Multiple Testing: Address the challenges of multiple hypothesis testing in economic research.

Based on the Book

isle is directly inspired by and closely follows:

It is highly recommended to use isle in conjunction with the ISLP textbook for a comprehensive understanding of the theoretical concepts and their practical implementation.

The notebooks in isle adapt and extend the Python labs from ISLP, providing an economics-specific perspective.

Notebooks

The repository is organized into chapters, mirroring the structure of the ISLP book. Each chapter contains:

  • Jupyter Notebooks: Interactive notebooks with code, explanations, and exercises.
  • Economic Examples: Data sets and examples relevant to economic analysis.

You can navigate the notebooks using the three-stripe menu button in the upper-left corner on mobile devices or the table of contents panel on the left side of the browser window.

Getting Started

  1. Run in your browser: The notebooks are designed for in-browser execution using platforms like Binder or JupyterLite. Click on the links to launch the interactive environment.
  2. Run & Experiment: Execute the code cells, modify parameters, and observe how the results change.

Contributing

We welcome contributions! If you'd like to improve isle by:

  • Fixing bugs
  • Adding new notebooks or exercises with an economics focus
  • Improving existing content
  • Suggesting enhancements

Please follow these steps:

  1. Fork the repository on GitHub.
  2. Create a new branch for your changes.
  3. Make your changes and commit them with clear, concise messages.
  4. Submit a pull request to the main branch of the isle repository.

License

This project is licensed under the MIT License - see the LICENSE file for details.

"An Introduction to Statistical Learning" and its associated materials have their own licensing terms, which should be respected.

Acknowledgements

  • Gareth James, Daniela Witten, Trevor Hastie, Rob Tibshirani, and Jonathan Taylor for their outstanding book, "An Introduction to Statistical Learning."
  • The developers of the Python libraries used in isle, including statsmodels, scikit-learn, pandas, and numpy.
  • The Executable Book Project for developing MyST and Thebe.
  • The open-source community for their invaluable contributions to statistical learning and data science.

Embark on your statistical learning journey with isle and discover its power in the realm of economics!