FairLanka

BEng (Hons) Software Engineering | Final Year Project

Rashmini Chamathsara Atthanayake

Fake news is spreading rapidly across various online platforms, and it is no longer confined to text-based content. In the modern digital era, social media has emerged as the primary channel for the dissemination of fake news. Many individuals tend to believe such misinformation and are easily deceived by it. Therefore, it is essential to develop a strong multimodal fake news detection system.Existing research and detection system only has focus on English language content. However,fake news is not limited to English. It is also produced and circulated in low-resource languages such as Sinhala, in countries like Sri Lanka. Limited attention has been paid to addressing fake news in these languages.This study proposes the development of multimodal model employing Machine Learning and Natural Language Processing techniques to detect fake news in both Sinhala and English. The proposed system will process a large cross lingual dataset using preprocessing techniques such as feature scaling and outlier detection. In addition, text extraction methods will be utilized to identify and analyze textual information from memes and other social media content. The model will subsequently be trained and evaluated using supervised learning methods to optimize fake news detection accuracy.