Abstract:

Abstract Over the last few years, the rise of fake news on social media has emerged as a significant issue, posing a potential threat to individuals, organizations, and society as a whole. As a solution to this issue, researchers have been using various natural language processing (NLP) techniques to detect fake news. In this study, we introduce a new strategy for fake news detection and classification. Our approach involves enhancing the performance of accuracy through fine-tuning, by merging BEART model with the proposed model DCNN. We have collected the data from secondary sources and combined it into a unified dataset. To improve its quality, we performed various processes such as data cleaning, transformation, integration, and reduction, which involved techniques like stop word removal, tokenization, and stemming, resulting in binary classification. Therefore, DCNN" was trained to classify news articles as real or fake, and the experiments on the dataset show that this approach performs better than several recent studies for detecting fake news, achieving high accuracy

Keywords: