Authors: Pol Pastells Vilà, Narcís Font Massot (December 2020)
Explainable ML project for a master's course (see project_description.pdf) using XGBoost and Shap.
Predicting the GDP Growth for the next year using the World Development Indicators with SQL from Kaggle:
https://www.kaggle.com/mariapushkareva/world-development-indicators-with-sql
Actually, we used a slightly different version, with data only till 2010, and the objective was to predict the 2011 GDP growth.
For the program to work, unzip the database and leave it in the same folder as cli.py.
We used a virtualenv to run the program, all the dependencies can be found inside requirements.txt.
The folders are structured as follows:
-
Analysis: contains notebooks to document the process and the final model explained.
-
Logs: folder where logging saves the logs of the program.
-
Models: folder to save the model and created plots.
-
Reports: html version of the notebooks.
-
Utils: modules called from
cli.py(main program)