Skip to content

GraphicsMonster/Fraudulent-transactions-XGBoost

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Fraudulent Credit Card Transaction Detection using XGBoost

Overview

This Python notebook demonstrates a fraudulent detection model developed for a dataset containing credit card transaction details. Similar to a previous implementation using neural networks, this model uses XGBoost, a popular gradient boosting library. The dataset features were anonymized due to regulatory requirements. This model achieved an accuracy of 99.8% on the unseen test set, which is an improvement on the neural nets implementation that achieved 97.38% accuracy.

Model Architecture

The XGBoost model is known for its ensemble learning capabilities and is particularly effective in tabular data scenarios. The model is configured with default hyperparameters, and no extensive tuning is performed due to the impressive out-of-the-box performance.

Features

  • Anonymized Feature Engineering: The model showcases the capability to engineer relationships between anonymized features, demonstrating the ability to learn patterns without explicit knowledge of individual feature meanings.

  • Tabular Data Handling: XGBoost excels in handling tabular data, making it suitable for credit card transaction datasets.

  • High Accuracy: The model achieved a high accuracy of 99.8% on the unseen data(test set).

Dependencies

  • Python 3.x
  • PyTorch
  • pandas
  • scikit-learn

Ensure the required packages are installed using:

pip install torch pandas scikit-learn

Result

  • Test Set Accuracy: 99.8%
  • Can't test the model on a validation set since the data is anonymized and hence it's difficult to synthesize a validation set.

Acknowledgments

License

This project is licensed under the MIT License.

About

XGBoost model trained on anonymized features to detect fraudulent credit card transactions. Dataset from Kaggle.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published