This project is a machine learning-based solution to detect fraudulent financial transactions using models like Random Forest, Logistic Regression and XGBoost. It includes preprocessing, scaling, model training, and a basic web interface to test predictions on transaction data.
Fraudulent-Transactions-Detection/
├── static/
│ └── style.css
├── Templates/
│ ├── index.html
│ ├── input.html
│ └── output.html
├── Train_data/
│ ├── fraudTrain.csv
│ └── fraudTest.csv
├── app.py
├── fraud.ipynb
├── model.pkl
├── scaler.pkl
- Logistic Regression
- XGBoost Classifier
- Random Forest
These models are trained on labeled transaction data to distinguish between legitimate and fraudulent transactions.
The dataset is stored in the Train_data/ folder:
fraudTrain.csv– training datafraudTest.csv– testing/validation data
Each dataset includes transaction features such as amount, category, merchant info, and a label indicating fraud (1) or not (0).
- Python
- Pandas, NumPy
- Scikit-learn
- XGBoost
- Flask (for web deployment)
- HTML/CSS (used in Templates and static folders)
- Clone the repository
git clone https://github.com/JayanthSrinivas06/Fraudulent-Transactions-Detection.git
cd Fraudulent-Transactions-Detection- Install dependencies
pip install -r requirements.txt- Run the Flask app
python app- Visit the web interface
Open your browser and go to: http://localhost:5000
Both models are evaluated using metrics like:
- Accuracy
- Precision & Recall
- Confusion Matrix
Random Forest generally provides better accuracy and recall for fraud detection tasks.
Developed by Jayanth Srinivas Bommisetty
For questions or suggestions, feel free to reach out via LinkedIn
⭐ Star this repo if you found it useful!