Arabic Fake News Detection Using CNN and Fuzzy Logic

The distribution and dissemination of information face a major challenge due to the fast spread of fake news in digital communications and social media. Most International governments, in general, and Arab governments, in particular, were keen to impose sanctions on the crime of spreading rumours. Due to its linguistic and cultural difficulties, existing methods for detecting fake news perform poorly when applied to Arabic. The need to find a solution prompted this research that will use convolutional neural networks (CNNs) and fuzzy logic to provide a new approach to solving this problem. The proposed model combines the power of convolutional neural networks for feature extraction with the ability of fuzzy logic for accurate classification in order to adapt to the dynamic nature of misinformation in the Arabic language. The research aims to create a more reliable digital information environment by improving accuracy by considering the nuances of the Arabic language and cultural context. The model will be trained and evaluated using the Arabic Fake News Dataset (AFND) to ensure it accurately represents the nuances of the Arabic language. The process begins by extracting features of convolutional neural networks and then applying fuzzy logic for more accurate classification. An accuracy of up to 90% was achieved in this research. By advancing the field of Arab fake news identification, this study hopes to improve Arab society’s access to reliable and reputable digital information.