This repository contains the code for the course "Machine Learning in Econometrics" at the Erasmus University Rotterdam. The code is written in Python. The code is created by the "free jacamar of mastery" group by Sem van Embden, Anna Grefhorst, Jaap Jansen & Luuk Omvlee.
Make sure to install the required packages by running the following command:
pip install -r requirements.txt
Additionally, to ensure PyTorch installs with CUDA support, you can specify it when running the installation command manually, as PyTorch installation with CUDA versions depends on your specific system configuration. For example, to install PyTorch with CUDA 11.8, you can use the following command:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
You can replace cu118 with the appropriate CUDA version for your system (e.g., cu117, cu116). If CUDA is not available, PyTorch will fall back to the CPU version.
For the final submission of the assignment, only the XGBOOST model (part 4) in main.ipynb
is used. The code can be run to train a new model, or a saved model can be loaded.