Digital Transaction Fraud Detection System (C++)

Enterprise-grade, rule-based fraud detection engine for digital payment systems. Built in pure C++ with a modular architecture, persistent storage, and compliance-style controls.

---

🚀 Overview

This project simulates a real-world fintech fraud monitoring system similar to those used by banks and digital wallets. It focuses on deterministic, rule-based fraud detection with enterprise controls such as:

• Risk scoring
• Automated account blocking
• Persistent audit trails
• Admin fraud monitoring
• Credit/Debit transaction flows

The project is developed with scalable design principles aligned with Accenture-style enterprise delivery.

---

✅ Key Features

v1.0 – Core Rule-Based Fraud Detection (Stable)

• User account creation & secure login
• Debit-only transaction engine
• Rule-based fraud risk scoring (0–100)
• Automated account blocking
• File-based persistent storage (.dat)
• Case-insensitive geo-location validation

v1.1 Phase 1 – Enterprise Transaction Foundation

• Credit + Debit transaction classification
• Fraud logic applied only to DEBIT transactions
• Upgraded audit log format with transaction type
• Enterprise-ready transaction object model

v1.1 Phase 2 – Wallet Recharge & Trusted CREDIT Flow

• Wallet recharge functionality
• CREDIT transactions increase wallet balance
• CREDIT transactions bypass fraud checks
• Dual-entry audit logging (CREDIT & DEBIT)

v1.1 Phase 3 – Security & Compliance Controls

• Masked password input (***** style)
• Immediate forced logout on fraud detection
• Rolling 24-hour daily transaction compliance limit (₹1,00,000)

v1.1 Phase 4 – Admin Fraud Monitoring & Readable Audit Logs (Current)

• Admin login and monitoring dashboard
• View all users and blocked accounts
• View all transactions
• View suspicious (high-risk) transactions
• Human-readable timestamps for audit logs in admin dashboard

---

v1.1 Enterprise Edition Status

With Phase 4 completed, v1.1 now represents a full enterprise-grade fraud detection and compliance system with user, admin, security, audit, and regulatory controls integrated.

---

🧠 Fraud Detection Rules

Risk Score = Amount Risk + Frequency Risk + Geo Risk

Rule	Description	Risk
High Amount Spike	Debit > ₹50,000	+40
Frequency Burst	≥ 3 debits in 60 seconds	+30
Geo Mismatch	Non-INDIA location	+30

Blocking Threshold: Risk ≥ 70 → Account Blocked

---

🏗️ Architecture

FraudDetectionSystem/
│
├── src/
│   ├── main.cpp
│   ├── User.cpp
│   ├── Transaction.cpp
│   ├── FraudEngine.cpp
│   ├── FileManager.cpp
│
├── include/
│   ├── User.h
│   ├── Transaction.h
│   ├── FraudEngine.h
│   ├── FileManager.h
│
├── data/
│   ├── users.dat
│   ├── transactions.dat
│
├── README.md

---

⚙️ Technology Stack

• Language: C++17
• Paradigm: Object-Oriented Programming (OOP)
• Data Structures: map, vector
• Persistence: File Handling (fstream)
• Security Model: Rule-based risk engine
• Platform: Cross-platform (Windows / Linux)

---

🛠️ Build & Run

Prerequisites

• g++ compiler
• Git

Compilation

From the root directory:

g++ src/*.cpp -Iinclude -o fraud

Run

./fraud

(Windows: fraud or ./fraud)

---

🧪 How to Use

1. Create a new user
2. Login using credentials
3. Perform debit transactions
4. Observe fraud detection triggering
5. Recharge wallet (v1.1)
6. View blocked users as admin

---

🗃️ Data Storage Format

users.dat

userId password balance blocked

transactions.dat

userId amount location type timestamp

---

🔐 Security & Compliance Model

• Deterministic rule-based fraud detection
• Debit-only risk exposure
• Trusted credit inflow model
• Daily transaction ceiling
• Admin-level fraud visibility
• Full audit-log traceability

---

📈 Roadmap

• SQL database integration (MySQL/PostgreSQL)
• REST API layer (microservice-ready)
• Password hashing & encryption
• Admin export reports (CSV)
• Kafka-based transaction streaming

📄 License

This project is released under the MIT License for academic and demonstration purposes.