Accurately predicting stock prices is crucial for investment and risk management, but the non-stationarity of the financial market and the complex correlations among stocks pose challenges to traditional models (ARIMA, LSTM, XGBoost), resulting in difficulties in effectively capturing dynamic patterns and limited prediction accuracy. To this end, this paper proposes the Financial Spatio-Temporal Graph Attention Network (FSTGAT), with the following core innovations: temporal modelling through gated causal convolution to avoid future information leakage and capture long- and short-term fluctuations; enhanced spatial correlation learning by adopting the Dynamic Graph Attention Mechanism (GATv2) that incorporates industry information; designing the Multiple-Input-Multiple-Output (MIMO) architecture of industry grouping for the simultaneous learning of intra-group synergistic and inter-group influence; symmetrically fusing spatio-temporal modules to construct a hierarchical feature extraction framework. Experiments in the commercial banking and metals sectors of the New York Stock Exchange (NYSE) show that FSTGAT significantly outperforms the benchmark model, especially in high-volatility scenarios, where the prediction error is reduced by 45–69%, and can accurately capture price turning points. This study confirms the potential of graph neural networks to model the structure of financial interconnections, providing an effective tool for stock forecasting in non-stationary markets, and its forecasting accuracy and industry correlation capturing ability can support portfolio optimization, risk management improvement and supply chain decision guidance.

On the one hand, the Internet of Medical Things (IoMT) in healthcare systems has emerged as a promising technology to monitor patients’ health and provide reliable medical services, especially in remote and underserved areas. On the other hand, in disaster scenarios, the loss of communication infrastructure can make it challenging to establish reliable communication and to provide timely first aid services. To address this challenge, unmanned aerial vehicles (UAVs) have been adopted to assist hospital centers in delivering medical care to hard-to-reach areas. Despite the potential of UAVs to improve medical services in emergency scenarios, their limited resources make their security critical. Therefore, developing secure and efficient communication protocols for IoMT networks using UAVs is a vital research area that can help ensure reliable and timely medical services. In this paper, we introduce a novel Stackelberg security-based game theory algorithm, named Stackelberg ad hoc on-demand distance vector (SBAODV), to detect and recover data affected by black hole attacks in IoMT networks using UAVs. Our proposed scheme utilizes the substantial Stackelberg equilibrium (SSE) to formulate strategies that protect the system against attacks. We evaluate the performance of our proposed SBAODV scheme and compare it with existing routing schemes. Our results demonstrate that our proposed scheme outperforms existing schemes regarding packet delivery ratio (PDR), networking load, throughput, detection ratio, and end-to-end delay. Specifically, our proposed SBAODV protocol achieves a PDR of 97%, throughput ranging from 77.7 kbps to 87.3 kbps, and up to 95% malicious detection rate at the highest number of nodes. Furthermore, our proposed SBADOV scheme offers significantly lower networking load (7% to 30%) and end-to-end delay (up to 30%) compared to existing routing schemes. These results demonstrate the efficiency and effectiveness of our proposed scheme in ensuring reliable and secure communication in IoMT emergency scenarios using UAVs.

Unmanned aerial vehicles (UAV) have emerged as prime technologies due to their compatible size and flexible architecture. UAV technology offers services in vast application such as inter-UAV communication, wireless sensors, and the future Internet of Things (IoT) due to its compatible architecture. A UAV’s speed varies while roaming, which may increase the risk of a connection failure. Various routing schemes have provided solutions to address this essential issue for three-dimensional (3D) UAV-based networks. The main category of UAV routing schemes is position-based routing schemes, which choose the best route based on the UAV’s location. However, position-based routing has the drawback that it depends on exact positioning and tracking. An efficient routing scheme can resolve the significant issue associated with UAV mobility in a 3D environment. This paper aims to address the issues of static preloaded location values by presenting a hybrid routing scheme named the Position-Monitor-based Hybrid Routing Protocol (PMHRP), which takes advantage of both geographic and topology-based routing protocols. The PMHRP establishes the shortest possible route based on a UAV’s Global Positioning System (GPS). Moreover, the proposed protocol utilizes the links for data forwarding. Furthermore, a disaster-based UAV scenario is adopted to provide connections to IoT devices. A detailed comparison analysis shows the proposed scheme’s extreme performance and results in up to 65% to 73% better packet delivery ratio (PDR) than batch mark schemes under standard 3D UAV scenarios. Compared to earlier work, the proposed scheme reduces the average delay by up to 68% to 75%. Further proposed routing schemes offer 70% to 72% more throughput than the existing routing schemes, and NRL (%) is 42% to 49% lower than the existing routing schemes. This happens because of the global routing information available at each UAV which is provided by the position head coordinator (PHC) UAV in the proposed work.

Anomaly anticipation in traffic scenarios is one of the primary challenges in action recognition. It is believed that greater accuracy can be obtained by the use of semantic details and motion information along with the input frames. Most state-of-the art models extract semantic details and pre-defined optical flow from RGB frames and combine them using deep neural networks. Many previous models failed to extract motion information from pre-processed optical flow. Our study shows that optical flow provides better detection of objects in video streaming, which is an essential feature in further accident prediction. Additional to this issue, we propose a model that utilizes the recurrent neural network which instantaneously propagates predictive coding errors across layers and time steps. By assessing over time the representations from the pre-trained action recognition model from a given video, the use of pre-processed optical flows as input is redundant. Based on the final predictive score, we show the effectiveness of our proposed model on three different types of anomaly classes as Speeding Vehicle, Vehicle Accident, and Close Merging Vehicle from the state-of-the-art KITTI, D2City and HTA datasets.

Since the article “Traffic Anomaly Prediction System Using Predictive Network” by Riaz et al.