Enhancing The Accuracy of Academic Document Plagiarism Detection Through Big Data and Similarity Measure Approaches

Debora Chrisinta (Universitas Timor)
Miko Purnomo (Universitas Timor)
Iis Aprianti (Universitas Timor)
Justin Eduardo Simarmata (Universitas Timor)

Journal of Innovation Information Technology and Application (JINITA)
Vol 8 No 1 (2026): JINITA, June 2026

This study develops a Big Data-based plagiarism detection system using Similarity Measures to enhance objectivity and efficiency in evaluating student assignments. Employing a Research and Development (R&D) approach, the system processes .pdf and .docx documents through text preprocessing and computes similarities using Jaccard and Cosine Similarity. Testing on assignments from the Information Technology Program at the University of Timor showed that Cosine Similarity is more sensitive to sentence structure, while Jaccard better detects identical phrases. Around 40% of documents exceeded a 40% similarity threshold, indicating potential plagiarism. The system offers fast, flexible detection via a web interface, though it remains limited in identifying semantic or paraphrased content. Future improvements will incorporate NLP techniques to enhance accuracy and academic integrity