FinGuard Enterprise - Bank Security

AI-driven financial fraud detection system with:

• Streamlit analyst dashboard
• XGBoost model inference
• Forensic PDF report generation
• Independent Android APK with offline transaction risk analysis

Features

• Real-time fraud transaction simulation and scoring
• Interactive forensics dashboard and network explorer
• Audit log tracking
• Fraud report PDF download
• Independent Android app (finguard-mobile.apk) with native, offline fraud scoring UI

Project Structure

Bank-Security/ ├── main.py ├── modules/ │   ├── data_loader.py │   ├── pdf_generator.py │   ├── risk_calculator.py │   └── ui_components.py ├── android/ │   └── app/... ├── fraudmodelxg.pkl ├── requirements.txt └── .github/workflows/android-release.yml

Setup (Python App)

• Install dependencies:

pip install -r requirements.txt

• Run:

python -m streamlit run main.py

FinGuard_Launcher.bat

Dataset Handling

• CSV datasets are included in version control for full local reproducibility.
• If CSVs are absent, the app still supports synthetic fallback demo data.
• Model file fraudmodelxg.pkl remains part of the repo for inference.

Android APK

• Source: android/
• Build output in release: finguard-mobile.apk
• The APK runs independently from Streamlit with local risk-score calculations.
• Streamlit workflow remains unchanged and continues to run from main.py.

Release Automation

• Tag push v* triggers Android build workflow:

License