Skip to content

High precision and recall rate using Over and Undersampling technique

Notifications You must be signed in to change notification settings

amayumradia/Kaggle_imbalanced_data_credit_card_fraud

Repository files navigation

Kaggle Dataset for imbalanced data of credit_card_fraud transcation

Highly imbalanced data was classifed using various classification techniques.

- Logistic Regression

With and without class weights

With and without Binarizing

- Undersampling and Oversamling using imblearn

- Logistic comparison with Random forest

About

High precision and recall rate using Over and Undersampling technique

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published