Plagiarism Detection using ML

Project Overview

This project is a web-based tool that detects plagiarism by comparing text against internet sources and between files. It uses Python and the KMP (Knuth-Morris-Pratt) algorithm for efficient text pattern matching, combined with machine learning techniques to identify potentially plagiarized content with high accuracy.

Key Features

• Dual detection approach:
  • Internet source comparison through Google Search API
  • File-to-file comparison using the KMP algorithm
• Interactive web interface built with Flask
• Detailed similarity reports with percentage matches
• Highlighted text sections showing potential plagiarism
• Support for multiple file formats (TXT, DOC, PDF)
• Fast processing with optimized text pattern matching

Video Preview

Technical Implementation

The system architecture includes several key components:

• Flask web framework for the frontend and API
• KMP algorithm implementation for efficient string matching
• Google Search API integration for web source comparison
• Text preprocessing with NLTK for tokenization and normalization
• Machine learning models for similarity detection
• Document parser for handling multiple file formats

Algorithm Highlights

The core of the plagiarism detection system uses the KMP algorithm, which:

• Performs linear-time string matching (O(n+m) complexity)
• Uses a preprocessing step to build a partial match table
• Efficiently identifies matching patterns without backtracking
• Works effectively with large documents and complex text patterns

Applications

This plagiarism detection tool is useful for:

• Educational institutions checking student assignments
• Publishers verifying original content
• Researchers validating citation accuracy
• Content creators ensuring originality