💳 Credit Card Fraud Detection using DBSCAN Clustering

An unsupervised machine learning approach to detect fraudulent credit card transactions using DBSCAN clustering algorithm.

🎯 Overview

Implements an unsupervised anomaly detection system for identifying potentially fraudulent credit card transactions without labeled data.

🚀 Installation

pip install pandas numpy matplotlib scikit-learn

💻 Usage

Open DBSCAN_clustering.ipynb in Jupyter Notebook or Google Colab and run all cells sequentially.

📊 Dataset

Works with credit card transaction datasets. Recommended: Kaggle Credit Card Fraud Detection Dataset

🔬 Methodology

1. Data Preprocessing — handles missing values and standardizes features
2. DBSCAN Clustering — density-based anomaly detection
3. Anomaly Detection — points labeled -1 are potential fraud
4. PCA Visualization — 2D projection of clusters

📚 References

• Ester, M., et al. (1996). “A density-based algorithm for discovering clusters”
• Scikit-learn DBSCAN Documentation