The rapid accumulation of large-scale, long-term meteorological data presents unprecedented opportunities for data-driven weather modeling and high-resolution numerical weather prediction. While various deep learning techniques—such as Long Short-Term Memory (LSTM), Recurrent Neural Networks (RNNs), and Graph Neural Networks (GNNs)—have been explored for weather forecasting, the complex spatial dependencies within historical meteorological data, particularly dynamic spatial correlations, remain insufficiently addressed. To tackle this challenge, we propose a Dynamic Spatio-Temporal Fusion Graph Network (DSTFGN), a novel module that integrates multivariate time-series analysis with graph-... More >

Traffic flow prediction is a critical component of Intelligent Transportation Systems (ITS) and smart city infrastructures. This survey paper provides a comprehensive analysis of recent advancements in deep learning-based approaches for traffic flow prediction, focusing on spatiotemporal correlations and attention mechanisms. We systematically review five seminal papers that propose innovative neural network architectures including DHSTNet, Att-DHSTNet, and ASTMGCNet for citywide traffic prediction. Our survey examines their methodologies, key contributions, experimental results, and comparative performance. We organize the discussion around three main themes: (1) modeling dynamic spatiotemp... More >

In the modern world, disseminating false information is a problem that must be addressed, and algorithms based on machine learning are used to spot and stop the spread of incorrect information. Due to the current unregulated development of false news fabrication and dissemination, democracy is continuously under threat. Fake news may mislead individuals while influencing them because of its persuasiveness and life sciences. Using data from the Web of Science, this study undertakes a bibliometric analysis of research on the application of machine learning for fake news identification. The research underscores the need for a streamlined approach to analyze data exclusively from the Web of Scie... More >