Utilization of NLP Techniques in Plagiarism Detection System through Semantic Analysis using Word2Vec and BERT

Detecting strongly paraphrased and translated texts is challenging for existing detection tools as they rely on traditional approach of word searching and matching. Automated systems play a crucial role in identifying instances of plagiarism, thereby upholding the integrity of intellectual work. The study presents a system that detects plagiarism in paraphrased texts using Natural Language Processing Techniques through Word Embedding Techniques, specifically Word2Vec and Bidirectional Encoder Representations from Transformers (BERT). By combining different approaches and techniques, the results showed that the hybrid model achieved 93% accuracy in detecting paraphrased plagiarism. After the integration of the model to the system, the evaluation using the ISO 25010 accumulated excellent results of Functional Suitability, Performance Efficiency, Usability, and Reliability with means of 4.29, 4.29, 4.59, and 4.34, respectively.