A hybrid deep learning and differential evolution approach for accurate fake news detection

The rapid spread of misinformation on digital platforms has led to widespread trust breakdowns, influencing critical decisions in politics, healthcare, and finance. Existing fake news detection methods often struggle with scalability, accuracy, and adaptability, necessitating more robust and efficient solutions. The proposed study combines deep learning with Differential Evolution optimization in a hybrid method to improve fake news detection accuracy and system efficiency. This study presents a novel hybrid approach combining deep learning techniques with Differential Evolution (DE) optimization to enhance the accuracy and scalability of fake news detection systems. The proposed framework integrates Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, and attention mechanisms for robust feature extraction and classification. By leveraging DE optimization, the model fine-tunes its parameters for improved convergence and detection performance. Experimental results on benchmark datasets demonstrate that the hybrid approach achieves superior accuracy, precision, recall, and F1-score compared to existing methods such as Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and standalone deep learning models. This research highlights the effectiveness of integrating advanced optimization techniques with deep learning to mitigate the challenges of misinformation dissemination and offers a scalable solution for real-time detection of fake news.