Machine Learning for Computational Economics (MLCE)

This repository hosts the public course website, lecture notes, slides, replication code, and Pluto notebooks for the course:

Machine Learning for Computational Economics
Instructor: Dejanir Silva (Purdue University)
Institution: EDHEC Business School

📘 Course website:
👉 https://dejanirsilva.github.io/mlce

---

📂 Repository Contents

Website Files

• index.html – Course homepage (generated via Quarto from index.qmd)
• Module0X/ – HTML slides and figures for each module (e.g. Module01/Module01_Slides.html)
• notebooks/ – Static HTML versions of the interactive Pluto notebooks
• Lecture_Notes_MLCE.pdf – Complete lecture notes for the course

Source Code

• src/Module0X/ – Julia source code for each module, including:
  • Example scripts and numerical experiments
  • Benchmarking and illustration code
  • DPI implementation routines
  • Pluto notebooks (files of the form NB_*.jl)

---

🎓 Course Slides

Each module has a full HTML slide deck (also linked from the course website):

• Module 01 – Introduction
  Module01/Module01_Slides.html

• Module 02 – Discrete-Time Methods
  Module02/Module02_Slides.html

• Module 03 – Continuous-Time Methods
  Module03/Module03_Slides.html

• Module 04 – Fundamentals of Machine Learning
  Module04/Module04_Slides.html

• Module 05 – The Deep Policy Iteration Method
  Module05/Module05_Slides.html

---

💻 Interactive Pluto Notebooks

Static HTML previews of the notebooks live in the notebooks/ folder:

• notebooks/NB_ThreeChallenges.html
• notebooks/NB_BlackScholes.html
• notebooks/NB_FittingDNN.html
• notebooks/NB_TwoTrees.html

The original Pluto notebooks are in src/Module0X/ and can be run locally.

To launch Pluto and open notebooks:

using Pluto
Pluto.run()

Or run a specific notebook directly, for example the Module 2 notebook:

pluto run src/Module02/NB_ThreeChallenges.jl

🔍 Replication Code

All replication code for figures, examples, and numerical methods discussed in the course is located under src/.

The codebase covers, among other things:

• Value Function Iteration (VFI)
• Policy Function Iteration (PFI)
• Endogenous Gridpoint Method (EGM)
• Deep Policy Iteration (DPI)
• Neural-network approximations to value and policy functions
• Automatic differentiation, stochastic calculus tools, and related utilities

The structure is modular so that users can reuse individual components in their own projects.

---

📄 License and Use

The material in this repository is intended for educational and research purposes.