Development of an Effective Hybrid Text Plagiarism Detection System using Machine Learning Techniques

In recent times, there has been a great spread of plagiarism as a result to the advancement on internet technology, which has brought about large volume of information to be share on the internet in several fields and discipline. The design and implementation of a Plagiarism Detection System for Nigerian Public University’ is a detailed approach to dealing with academic plagiarism at the university. The study examined the existing systems relating to the plagiarism detection model as well as system, designed the model for the plagiarism system, implement the proposed system, and evaluate the developed system on various performance metrics. The system used a Hybrid Methodology by combining text-based, semantic-based, and machine learning-based techniques to analyze academic submissions. The proposed system has been shown to be efficient and effective, with an average accuracy of 80%, precision score of 0.90, recall score of 0.80, F-measure of 0.82, granularity of 0.93(≈1.000) and a plagdet percent or score of 0.8856 in detecting plagiarized documents. The study highlights the risks associated with academic plagiarism, the development of a database of research papers, and the evaluation of machine-learning techniques for plagiarism detection. The report emphasizes the importance of developing effective plagiarism detection systems to promote academic integrity and originality in academic institutions. This project is a significant contribution to the computer science field and has the potential to positively impact academic integrity in Nigerian public University and beyond. Keywords: Hybrid Text, Plagiarism, Detection, Machine Learning and Techniques