Publication

Abstract : Public discourse and information integrity are seriously threatened in the digital era by the quick dissemination of false information and fake news in many languages. The fact that misleading information can appear in numerous languages in a diverse nation like India makes identifying fake news extremely difficult. The existing research on fake news detection depends on the length of the text. The proposed work overcomes this limitation by completely employing a simple Support Vector Classifier, which can classify text with distinct lengths and is cost-effective. Considering that English is an international language and Hindi is the national language of India, a bilingual fake news detection model has been developed in this work. The experimental findings vividly demonstrates that the proposed work achieved 89% accuracy. The model correctly predicted all the new text, regardless of its length and a deeper exploration of how the model handles semantic meaning of text is made. This makes the proposed system counter misinformation by handling real-time user input for fake news classification tasks.