Learning guide

Examples accompanying Estimating Nonlinear Heterogeneous Agent Models with Neural Networks by Hanno Kase, Leonardo Melosi, and Matthias Rottner.

Learn how to solve and estimate macroeconomic models with neural networks, starting with a small New Keynesian model, where each extension introduces one change at a time, and moving on to progressively more complex models.

Start with the overview. It explains the method step by step and shows how the examples are organized. Setup installs the Python environment.

Learning guide

The learning guide walks through the method and every example in detail. These pages follow it and connect its equations to the code.

Download the learning guide (PDF)

Examples

• New Keynesian model, with ten extensions
• One-asset HANK
• Krusell–Smith with one asset and two assets

Every example ships its trained networks and saved results, so the code can reproduce the figures without retraining.

Citation

Please cite the main paper when using the code or these examples.

Hanno Kase, Leonardo Melosi, and Matthias Rottner. Estimating Nonlinear Heterogeneous Agent Models with Neural Networks.

@article{kase_melosi_rottner_neural_networks,
  author = {Kase, Hanno and Melosi, Leonardo and Rottner, Matthias},
  title  = {Estimating Nonlinear Heterogeneous Agent Models
            with Neural Networks},
  url    = {https://www.hannokase.com/research/deep-learning-zlb/KMR_NeuralNetwork_Hank.pdf}
}