Invoice Fraud Detection System

Overview

The Invoice Fraud Detection System is a comprehensive tool designed to analyze invoices for potential fraud. It uses advanced techniques such as OCR (Optical Character Recognition), pattern matching, and risk scoring to identify anomalies in invoices. The system supports multiple formats, including PDFs and images, and provides detailed reports for review.

Features

• Invoice Analysis: Extracts and analyzes key details such as amount, date, supplier, and GST.
• Risk Scoring: Assigns a risk score based on various factors like supplier validity, amount patterns, and date anomalies.
• Trusted Supplier Tracking: Maintains a database of trusted suppliers and their transaction history.
• Monthly Reports: Generates detailed monthly reports summarizing processed invoices and flagged risks.
• Streamlit Dashboard: Provides an interactive web interface for uploading and analyzing invoices.

Project Structure

• feature3_real.py: Core logic for invoice processing, fraud detection, and risk analysis.
• app.py: Streamlit-based web application for user interaction and visualization.
• fraud_detection_report.txt: Example output report generated after processing invoices.
• supplier_history.db: SQLite database for storing supplier information and transaction history.

Installation

1. Clone the repository:

   git clone <repository-url>
   cd deployed-3rd-AI-feature

2. Install dependencies:

   pip install -r requirements.txt

3. Download the required Tesseract OCR engine:

   • Tesseract Installation Guide

4. (Optional) Download the SpaCy language model:

   python -m spacy download en_core_web_sm

Usage

Command-Line Interface

1. Run the main script to process invoices:

   python feature3_real.py

2. Follow the prompts to specify the directory containing invoices.

Streamlit Web Application

1. Launch the Streamlit app:

   streamlit run app.py

2. Open the provided URL in your browser.
3. Upload invoices or use the sample invoice feature for analysis.

Supported Formats

• PDF: Extracts text using PyPDF2.
• Images: Supports JPEG, PNG, and TIFF formats using Tesseract OCR.

Reports

• Risk Factors: Detailed breakdown of risk scores for each invoice.
• Monthly Summary: Aggregated statistics for processed invoices, flagged risks, and trusted suppliers.

Example Output

Risk Analysis

File: invoice_1.pdf
Risk Score: 0.39

Risk Factors:
  - amount: 0.00
  - date: 0.80
  - supplier: 0.20
  - gst: 0.80
  - line_items: 0.50

Details:
  Amount: $71214.00
  Date: Not found
  Supplier: Tech Corp (Registered Business)

Monthly Report

Monthly Analysis Report - 2024-12
==================================================
Total Invoices Processed: 76
Flagged for Review: 33
Total Amount Processed: ₹2,704,694.01
Trusted Suppliers: 0

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Submit a pull request with a detailed description of your changes.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• Tesseract OCR for text extraction.
• Streamlit for the interactive web interface.
• PyPDF2 for PDF text extraction.
• SpaCy for natural language processing.

Contact

For questions or support, please contact the project maintainer at u23ai052@coed.svnit.ac.in