Fraud Detection Model

Author: Huseyin Zengin

This project implements a machine learning pipeline for detecting fraudulent transactions using a Kaggle dataset.

Repository Structure

fraud_detection_xgboost.ipynb : Jupyter notebook containing the full workflow.

Methods

The notebook includes:

• Exploratory Data Analysis
• Missing value inspection
• Outlier detection
• Feature preparation
• Model training
• Model evaluation using ROC-AUC

Tools Used

• Python
• Pandas
• Scikit-learn
• XGBoost
• Jupyter Notebook

Dataset

The dataset is too large for GitHub and can be downloaded from Kaggle.

Place the following files in the project directory before running the notebook:

• train.csv
• test_share.csv