Logistic Regression Model using scikit-learn
This project aims to enhance transaction auditing with a Logistic Regression model to detect fraudulent activities 🚨. Designed for seamless integration into automatic workflows and autoflagging systems, this tool efficiently identifies suspicious transactions.
- Type of Transaction
- Current System Fraud Flag
- Transaction Step
- Note: 'Step' represents a unit of time in the transaction process (hours, days, etc.)
- Transaction Amount
- Old Balance of Origin Account
- New Balance of Origin Account
- Old Balance of Destination Account
- New Balance of Destination Account
Utilizing One Hot Encoding throughout its process, this model offers a swift and reliable method to flag transactions for further personal inspection. While it may not replace existing systems, it significantly aids in managing large volumes of transactions efficiently 🔄.
Explore the code and see how advanced fraud detection can be achieved! 🚀
Note: the dataset is too large to upload to the repository