Credit Card Fraud Detection using Ensemble Methods (AdaBoost and Majority Voting)

Overview

This project aims to detect credit card fraud using ensemble learning methods, specifically AdaBoost and Majority Voting. The dataset used contains transaction details, including features such as transaction amount, merchant, and customer information. The goal is to classify transactions as either legitimate or fraudulent.

Key Features:

• Ensemble Methods: Utilizes AdaBoost and Majority Voting to improve model performance.
• Data Preprocessing: Handles imbalanced data and applies techniques like SMOTE for oversampling.
• Evaluation: Performance metrics such as accuracy, precision, recall, and F1 score are computed.
• Visualization: Interactive graphs and plots to better understand the performance and model decision-making process.

---

Technologies Used

• Python
• Scikit-learn
• Pandas
• NumPy
• Matplotlib & Seaborn
• imbalanced-learn (for oversampling techniques like SMOTE)

---

Dataset

The dataset used is a simulated credit card transaction dataset, containing features that represent transaction details. The target variable indicates whether the transaction was fraudulent (1) or legitimate (0).

Note: For the sake of this project, we have used a publicly available dataset, such as the Kaggle Credit Card Fraud Detection Dataset.

---

Installation

Example Outputs Model Performance Here’s an example of what the evaluation output may look like:

Accuracy: 98.5% Precision: 92.7% Recall: 91.2% F1-Score: 91.9%

Clone the Repository

git clone https://github.com/yourusername/credit-card-fraud-detection.git
cd credit-card-fraud-detection