Enhancing Fake News Detection Using a Gated Recurrent Units Model

The spread of fake news has grown significantly in the information age. Fake news has the ability to distort public perception, influence opinions, and shape behaviors, thereby undermining trust and unity within society. As communities become increasingly exposed to misleading information, it can foster division, spread fear, and heighten tensions. Detecting fake news manually requires critical thinking, a skeptical mindset, and careful evaluation of the information at hand. Although machine learning can be used for fake news detection, it is a time-consuming process, especially during the training phase. Therefore, the implementation of deep learning is expected to enhance the accuracy of fake news detection. In this study, the Gated Recurrent Unit (GRU) is used to detect fake news. The findings suggest that the GRU model may be more effective in fake news detection compared to the commonly used Long Short-Term Memory (LSTM) model, with effectiveness measured using accuracy, precision, recall, and F1-score.