FraudShield

Hybrid AI-based Bank Message Fraud Detection System

---

About

FraudShield is a learning project focused on detecting fraudulent bank SMS/messages using a hybrid rule-based and machine learning approach.

---

Hybrid Detection Approach

Rule-Based Detection

• URL pattern analysis
• Suspicious domain checks
• Unicode / homograph attack detection
• Fraud-related keywords

Machine Learning Model

• TF-IDF vectorization
• Logistic Regression classifier
• Confidence threshold handling

---

Classification Labels

• SAFE
• SUSPICIOUS
• FRAUD

The system exposes a FastAPI backend and includes a lightweight web demo to demonstrate the end-to-end flow.

---

System Flow

---

How To Run

1. Clone Repository

git clone https://github.com/adityamandre25/FraudShield.git

cd FraudShield

2. Create Virtual Environment

Windows:

python -m venv venv venv\Scripts\activate

Mac/Linux:

python3 -m venv venv source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Start Server

uvicorn app.main:app --reload

Open in browser:

http://localhost:8000

---

Sample Messages For Testing

SAFE:

Amazon Pay: ₹50 cashback has been credited to your Amazon Pay balance.

SUSPICIOUS:

Action required to maintain your access.

FRAUD:

Dear User, your SBI KYC has expired. Update immediately at http://sbi-verify-secure.in to avoid account freeze.

---

Notes

• This project is intended as a learning and demonstration exercise.