FraudShield v1.0.0 Initial Release

FraudShield v1.0.0 — Initial Release

Release Date: April 2026
Platform: Windows 10 / Windows 11 (64-bit)

---

What's New

Complete 8-module AI fraud detection desktop application, released as a standalone Windows installer. No Python installation required.

Core System

• Full PyQt5 desktop application with dark and light theme support
• Session-based authentication with bcrypt password hashing and account lockout
• SQLite3 database with Fernet-encrypted account numbers
• Role-based access control: Admin and Analyst roles

AI & Machine Learning

• 6 trained ML models: Logistic Regression, Random Forest, Decision Tree, SVM, Isolation Forest, K-Means
• Soft-voting ensemble model combining all 4 supervised classifiers
• Hybrid OR-gate model: flags fraud if Ensemble or Isolation Forest detects it
• Risk scoring (0–100) on every transaction
• AI controls in Admin Panel: configurable threshold (60–90%), model retraining

Transaction Management

• Manual transaction entry with account number encryption
• Bulk CSV import supporting 100,000+ records
• Edit and review transactions with real-time fraud status

Alerts System

• Real-time alerts generated for transactions above risk threshold
• Severity levels: Low, Medium, High
• Sound notifications (WAV) with toggle control
• Alert resolution and audit trail

Analytics Dashboard

• Live stat cards: total transactions, fraud count, revenue, risk score
• Trend comparison vs. previous period
• Charts: fraud trend line, risk score distribution, monthly bar comparison
• Animated splash screen on startup

Reports & Export

• Professional PDF reports via ReportLab with header, summary table, and charts
• Excel export via openpyxl
• Date range and status filters
• Downloadable directly from the Reports tab

Admin Panel

• User management: create, view, delete users
• System activity logs and login history
• Database backup and restore
• AI model controls and threshold configuration

---

System Requirements

	Minimum	Recommended
OS	Windows 10 (64-bit)	Windows 11 (64-bit)
RAM	4 GB	8 GB
Disk	2 GB free	4 GB free
Python	Not required	Not required

---

Installation

1. Download FraudShield_Setup.exe from the assets below
2. Right-click → Run as administrator
3. Follow the setup wizard (default install: C:\Program Files\FraudShield)
4. Launch from the Desktop shortcut or Start Menu → FraudShield

---

Default Login Credentials

Username: admin
Password: admin123

⚠️ Change the default password immediately after first login via Profile → Change Password.

---

Key Capabilities

• AI fraud detection result in under 3 seconds per transaction
• Supports 100,000+ transaction records without performance degradation
• Fully offline — no internet connection required after installation
• Dark and light theme, switchable per user account
• All sensitive account numbers encrypted at rest with Fernet symmetric encryption

---

Known Limitations

• Windows only — macOS and Linux are not supported in this release
• Large CSV imports (50,000+ rows) may take 2–3 minutes to process
• First model training after a fresh database requires 5–10 minutes
• Sound notifications use winsound and require a Windows audio device

---

Files in This Release

File	Description
FraudShield_Setup.exe	Windows installer — run this to install

---

Team

Name	Contribution
Zainab	AI Model Development & Backend
Aqsa Nadeem	Database Design & Desktop UI
Zainab Minahil	Reporting, Integration & Deployment

Project Advisor: Prof. Muzammil Sadiq
University: University of Central Punjab, Lahore
Group ID: G1F22FYPCS011
Academic Year: 2025–2026

---

This software was developed as a Final Year Project for academic purposes only. All rights reserved © 2026.