Sosyal medyada sahte haber tespiti için durum sınıflandırması

In today's digital landscape, the availability of a diverse range of resources has accelerated the growth of news dissemination. This has enabled individuals to share their opinions through social media, sometimes using automated bots or fake user. Fake News Detection is a classification task which aims to determine whether news is credible or not. In recent years, there has been a growing interest in FND in Arabic languages, and many detection approaches have demonstrated the ability to identify fake news across various datasets. In our research, we combined word embedding techniques and a hybrid deep learning model, with optimizing preprocessing techniques. We also utilized feature extraction and selection to obtain accurate data from news articles. Experiments conducted on the publicly available AFND dataset have revealed impressive performance using the methods we applied. Our findings showcase the performance of deep learning models, including LSTM, BiLSTM, and CNN-BiLSTM, in comparison to traditional machine learning approaches like Logistic Regression, linear SVM, Naïve Bayes, and XGBoost. Notably, the CNN-BiLSTM model achieved the highest accuracy result, reaching 95%.