An Improved FakeBERT for Fake News Detection

• DOI: 10.2478/acss-2023-0018
• Publication Date: 2024-01-29T21:10:48Z
• Authors: Arshad Ali; Maryam Gulzar
• Keywords: Fake News

Abstract In the present era of internet and social media, way information dissemination has changed. However, due to rapid growth in amount news generated regularly unsupervised nature fake turns out be a big problem. Fake can easily build false positive or negative perception about person, an event. was also used as tool by propagandists during Coronavirus (COVID-19) pandemic. Thus, there is need use technology tag prevent its dissemination. Previously, different algorithms were designed detect but without considering semantic meaning long sentence dependence. This research work proposes new approach detection context COVID-19. The suggested uses combination Bidirectional Encoder Representations from Transformers (BERT) for extracting sentences, SVM pattern identification better COVID-19 dataset, evolutionary algorithm called Non-dominated Sorting Genetic Algorithm II (NSGA-II) distribute text Support Vector Machine (SVM) classification. improves accuracy 5.2 % removing certain ambiguity sentences.