🕵️‍♂️ Fraud Detection Model

This project uses machine learning to detect fraud in a labeled dataset. It covers data preprocessing, model training, evaluation, and visualization of fraud prediction.

---

📂 Project Structure

Fraud-Detection-Model/
├── data/
│   └── fraud_data.csv              # Dataset (CSV file)
├── images/
│   └── gridsearch_heatmap.png     # Heatmap of hyperparameter tuning
├── notebooks/
│   └── fraud_assignment.ipynb     # Jupyter notebook (main assignment)
├── fraud_assignment.py            # Converted Python script version
├── requirements.txt               # Project dependencies
├── README.md                      # Project overview (this file)
└── LICENSE                        # MIT License

📊 Dataset

The dataset (fraud_data.csv) contains observations labeled as fraudulent or not. The goal is to predict fraud using classification models.

---

🌟 Features

• Load and explore the dataset
• Train/test split
• Use of DummyClassifier as baseline
• Train classifiers like:
  • Logistic Regression
  • Support Vector Machines
  • Decision Trees
• Evaluate using:
  • Accuracy
  • Recall
  • Precision
  • Confusion Matrix
• Visualization using seaborn/matplotlib
• GridSearchCV for hyperparameter tuning

---

🛠️ Installation

Install dependencies:

pip install -r requirements.txt

Run the notebook:

jupyter notebook notebooks/fraud_assignment.py

Or run the Python script in PyCharm or terminal:

python fraud_assignment.py

---

📊 Sample Visuals

🔥 Grid Search Heatmap

---

📌 Requirements

- Python 3.7+
- scikit-learn
- pandas
- numpy
- matplotlib
- seaborn

---

📄 License

This project is licensed under the MIT License.

---

👤 Author

Mukesh Thenraj
📧 mukeshthenraj@gmail.com