Upgrading text clustering approach with optimized algorithm Makhija Mukta1 1Asst. Dean - IT, Professor, Integrated Academy of Management and Technology, Ghaziabad, Uttar Pradesh, India Online published on 30 August, 2024. Abstract Paraphrase detection performs crucial work in identifying plagiarism and ensuring content originality. This research proposes an innovative approach for paraphrase detection by employing sentiment analysis in conjunction with the BERT model. The BERT model, renowned for its contextual understanding capabilities, encodes pairs of sentences and generates comprehensive representations. This approach uses sentiment analysis to leverage the emotional nuances embedded within paraphrases. These emotional cues will provide valuable output for distinguishing between genuine paraphrases and plagiarised content. To check the efficiency of the proposed technique, vast practicals are conducted on benchmark datasets. The results demonstrate that the combination of sentiment analysis and the BERT model significantly improves the performance of paraphrase detection. The approach achieves high accuracy, precision, and recall levels, outperforming existing state-of-the-art methods. This research contributes to plagiarism detection by introducing a novel methodology that harnesses sentiment analysis and the BERT model. The findings underscore the significance of considering emotional cues in paraphrase detection, leading to more accurate identification of potential instances of plagiarism. These advancements promise to enhance content originality assessment in diverse domains, including academia, journalism, and online publishing. Top Keywords BERT Model, Data Sets, Sentiment Analysis. Top |