Skip to content

dynamo-pentester/Credit_card_fraud

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Credit_card_fraud

Fraud Detection Using Hybrid Deep Learning Model

Overview

This project focuses on detecting fraudulent transactions using a hybrid deep learning approach. It combines a Multi-Layer Perceptron (MLP) and an Autoencoder to improve fraud detection performance.

Dataset

Project Structure

📂 fraud-detection │── 📜 fraud_detection.ipynb # Jupyter Notebook with the complete workflow │── 📜 README.md # Project documentation │── 📜 requirements.txt # List of required dependencies

Installation

Clone the repository and install dependencies:

git clone https://github.com/DYNAMO_PENTESTER/fraud-detection.git cd fraud-detection pip install -r requirements.txt Steps in the Notebook Data Preprocessing

Load dataset and remove unnecessary columns

Encode categorical variables

Normalize numerical features

Handle class imbalance using SMOTE

Model Training

MLP Model: A deep learning model trained with fraud labels

Autoencoder: An unsupervised model trained on normal transactions to detect anomalies

Hybrid Model for Fraud Detection

Combines predictions from MLP and Autoencoder for better fraud detection

Evaluation & Visualization

Performance metrics: Accuracy, Confusion Matrix, ROC AUC Score

Feature importance analysis using SHAP

Precision-Recall curve for model evaluation

Results The hybrid model improves fraud detection accuracy by combining two approaches.

Precision-Recall curve analysis shows improved performance over a standalone MLP.

How to Run Execute the Jupyter Notebook step by step:

jupyter notebook fraud_detection.ipynb

Future Improvements:

Fine-tuning hyperparameters for better results

Exploring other anomaly detection techniques

Enhancing feature engineering for better fraud detection

🚀 Author Dynamo

About

Built a hybrid MLP–Autoencoder model for high-accuracy credit card fraud detection using imbalanced data handling and anomaly learning.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors