International Research Journal of Education and Technology

Plagiarism has become a significant concern in both academic and professional environments due to the widespread availability of digital content and the ease of copying and distributing it. Traditional manual detection methods are time-consuming, error-prone, and often fail to capture nuanced similarities across documents. This work presents a web-based Plagiarism Detection System that automatically identifies textual similarity between documents and helps maintain content originality.

The system is implemented using Python and the Flask framework for backend processing, with HTML, CSS, and JavaScript for frontend development. Users can securely register, log in, upload documents, and evaluate them for plagiarism. Input text is preprocessed through normalization, tokenization, and noise removal, improving detection accuracy. Multiple similarity detection techniques, including cosine similarity, Jaccard similarity, and n-gram matching, are applied to compare the uploaded document against a reference corpus.

The application generates a detailed similarity report with risk classifications such as High Risk, Medium Risk, Low Risk, and No Risk, allowing users to easily interpret results. Security measures, including authentication and password protection, ensure user data privacy. The system underwent functional, integration, security, and performance testing, demonstrating robustness, reliability, and efficiency. Future extensions include AI-based semantic similarity using transformer models and cloud-based deployment for scalability.

Keywords: Plagiarism detection, similarity analysis, Flask, Python, web application, content originality, document preprocessing.