sheromon / Focal_Loss_Keras

Multi-class classification with focal loss for imbalanced datasets

Multi-class classification with focal loss for imbalanced datasets

This repository was forked from Tony607's repository. I've made many changes, but thank you to Tony607 for putting together the initial Jupyter notebook!

How to Run

Clone or download this repo

git clone https://github.com/sheromon/Focal_Loss_Keras

Install required libraries

pip install -r requirements.txt

Download the fraud data set from Kaggle, extract the csv file, and put it in a directory named input in the main project directory. https://www.kaggle.com/ntnu-testimon/paysim1

From a terminal in the main project directory, run

jupyter notebook

The example baseline and focal loss code is in the notebook

src/fraud_detection_w_focal_loss.ipynb

Happy coding! Leave a comment if you have any questions.