This repository contains a project made for the Techonologies for Information Systems course.
The project consisted in the implementation of a bias-mitigation method (Resampling) on a machine learning model in order to correct the bias it presented when working with a given dataset.
The repo contains the following files:
- diabetes.csv: the dataset used to train and test the machine learning model
- healthcare.ipynb: the notebook containing the code
- requirements.txt: the required python modules
The function used in the code to apply the Resampling technique can be found in the following Repository: DALEX
The documentation for the function can be found here
The technique that the Repository implements has been taken from the following paper: Data preprocessing techniques for classification without discrimination
Lecturer: Prof. Letizia Tanca
Project Supervisors: Chiara Criscuolo, Tommaso Dolci
Authors: Giovanni De Lucia, Alessandro Ferri