Anomaly-Detection-in-Financial-Payment-Services

Dataset used: https://www.kaggle.com/datasets/ealaxi/paysim1

PaySim is a synthetic dataset which simulates real transactions using data collected by a multinational company in an African country.

Objectives

• Solving imbalanced dataset classification problem in detecting fraudulent transactions by balancing the majority and minority classes through the use of resampling techniques.
• Cleaning PaySim generated synthetic dataset for money transactions.
• Using resampling methods (undersampling, oversampling using SMOTE) and ensemble learning (XG-boosting).