GANerator

An MVP for applying Machine Learning (ML) in the credit risk modeling lifecycle.

It currently contains two types of models:

1. 🎯 Credit Stress Predictor

• ML Method: Random Forest - A machine learning algorithm that combines multiple decision trees to create a singular, more accurate result.

2. 🔄 Data Synthesizer

• ML Method: GAN (Generative Adversarial Network) - A deep learning method in which two neural networks compete with each other in a game, learning to generate new data with the same statistics as the training set

Installation

NOTE: The whole data set can be downloaded from Kaggle. Store the files application_test.csv and application_train.csv in the /data directory.

Create a new virtual environment (venv) and activate it:

python -m venv venv
source venv/bin/activate

Install requirements and the repo itself:

pip install -r requirements.txt
pip install -e .

Create a new jupyter kernel:

pip install ipykernel
python -m ipykernel install --user --name ganerator-dev

Usage

1. Credit Stress Predictor - Run the Risk Modelling Notebook.

2. Data Synthesizer - Run the Data Synthesizer Notebook.