Unveiling Deception: Different Strategies for Combatting Arabic Fake News

the proliferation of fake news has become a widespread problem in the contemporary era of information, posing possible harm to public discussions, political stability, and community confidence. To tackle this difficulty, it is necessary to develop strong and effective detection methods that can accurately recognize misinformation in several languages, including Arabic. The aim of this study is to create and assess different models specifically designed for the identification of fake news in Arabic language.Several of deep Learning models like Recurrent Neural Networks (RNN), Long short-term memory (LSTM), Bidirectional long short term memory (Bi-LSTM) and convolutional neural network (CNN) are used in this study. On the other hand, some of the traditional machine learning models like k-nearest neighbors (KNN), Random Forests (RF), Extreme Gradient Boosting (XGBoost), decision tree (DT) and multi-layer perceptron (MLP) are also used. Also, transformers, like Generative pre-trained Transformer 3 (Gpt-3) and Bidirectional Encoder Representations from Transformers (BERT), are used as a feature extraction to convert text into numerical representation and assign different number to each token. All of these models were used in evaluation which of them is the best model in the process of fake news detection for the Arabic Language Dataset.The findings demonstrate strong performance across deep learning models, with CNN-LSTM, Bi-LSTM and CNN-Bi-LSTM exhibiting the best accuracy rates of 93.2% ,92.6%, 93.6% respectively. The LSTM model achieves a moderate accuracy of 87.4%, whilst the RNN model demonstrates a more modest accuracy of 56.8%. The results emphasize the capability of deep learning methods in reducing the dissemination of fake information in Arabic-language datasets.