In the stock market, predicting the trend of price series is one of the most widely investigated and challenging problems for investors and researchers. There are multiple time scale features in financial time series due to different durations of impact factors and traders’ trading behaviors. In this paper, we propose a novel end-to-end hybrid neural network, a model based on multiple time scale feature learning to predict the price trend of the stock market index. Firstly, the hybrid neural network extracts two types of features on different time scales through the first and second layers of the convolutional neural network (CNN), together with the raw daily price series, reflect relatively short-, medium- and long-term features in the price sequence. Secondly, considering time dependencies existing in the three kinds of features, the proposed hybrid neural network leverages three long short-term memory (LSTM) recurrent neural networks to capture such dependencies, respectively. Finally, fully connected layers are used to learn joint representations for predicting the price trend. The proposed hybrid neural network demonstrates its effectiveness by outperforming benchmark models on the real dataset.

Weyl materials, exhibiting topologically nontrivial touching points in band dispersion, are fascinating subjects of research, which have been extensively studied in electron-related systems. In this paper, by employing first-principles calculations of topological phonons in wurtzite-structured phases of $MX$ chalcogenides (where $M=\mathrm{Zn}$ and Cd and $X=\mathrm{O}$, S, Se, and Te), we demonstrate the existence of ideal type-II Weyl phonons in wurtzite ZnSe, a well-known II-VI semiconductor. There are in the ${\mathbf{q}}_z=0.0$ plane six pairs of Weyl points stemming from the inversion between the two optical branches. The nontrivial phonon surface states and surface arcs projected on the semifinite (0001) surfaces are investigated. Phonon surface arcs connecting each pair of Weyl points with opposite chirality, guaranteed to be $0.55{\AA }^{-1}$ and very long, are readily observable in experiment. The opposite chirality of Weyl points with quantized Berry curvature produces the Weyl phonon Hall effect. Our results propose a potential platform for future experimental study of type-II Weyl phonons in realistic materials.

High-precision measurement of permittivity and thickness enables advanced quality control of microwave substrates. In this paper, a sub-wavelength planar microwave sensor is proposed, which is composed of a single-microstrip transmission line with a pair of coupled circular complementary spiral resonators. Thus, two robust band-stop notches in the transmission response are developed, which can be utilized for simultaneous thickness and permittivity measurements of a dielectric sheet. The mutual capacitance and inductance, which quantify the coupling effect between the two resonators, are added to the equivalent lumped circuit model. They enhance the frequency shifts, thus improving sensitivity. The results of full-wave simulations and measurements on five different types of samples verify the concept effectiveness.

Vehicle trajectory prediction (VTP) is an important task that can enhance the safety and efficiency of autonomous driving. However, existing VTP methods often struggle to fully extract spatiotemporal features, resulting in inaccurate prediction results. To solve this problem, we propose a hierarchical long short-term memory and graph attention network (HLSTM-GAT) model. First, we design a hierarchical network architecture to model different spatiotemporal features. The first-layer network (FLN) focuses on short-term trajectory information and immediate vehicle interactions to generate preliminary candidate trajectories. In the FLN, long short-term memory (LSTM) encoder processes historical trajectories to extract temporal features of vehicles, while a graph attention network (GAT) handles the LSTM-encoded outputs to capture spatial features between vehicles. Second, the second-layer network (SLN) combines the candidate trajectories generated by the FLN with historical trajectories to form comprehensive trajectories for accurate prediction. Subsequently, SLN adopts a GAT to process these comprehensive trajectories to precisely model the spatial relationships between vehicles. Moreover, we propose two distance threshold-based dynamic GAT models to perfectly capture spatial features between vehicles. They construct vehicle spatial interaction relationships at two distinct levels based on real-time distances in the FLN and predicted future distances in the SLN, respectively. These dynamic GAT models effectively filter out irrelevant information and enable the proposed method to consider potential future interactions. Finally, we conduct extensive experiments on two publicly available datasets, and experimental results demonstrate that the proposed method outperforms other state-of-the-art VTP methods in terms of prediction accuracy, robustness and computational efficiency.

Relay selection has been considered as an effective method to improve the performance of cooperative communication. However, the channel state information (CSI) used in relay selection can be outdated, yielding severe performance degradation of cooperative communication systems. In this study, the authors investigate how to select relays under outdated CSI in an amplify‐and‐forward cooperative communication system to improve its outage performance. The authors adopt maximum a posteriori (MAP) estimation to predict the actual signal‐to‐noise ratio (SNR) of each relay during data transmission and propose a single‐relay selection (SRS) strategy based on the MAP estimation (SRS‐MAP). To reduce the computational complexity, we approximate the a posteriori probability density of SNR and obtain a closed form of the predicted SNR. Simulation shows that SRS‐MAP outperforms the relay selection strategies given in the literature. In order to further improve the outage performance, a new multiple‐relay selection (MRS) method for outdated CSI is proposed, and by applying it to existing SRS strategies the corresponding MRS strategies are obtained. The authors find that the MRS strategies perform much better than their corresponding SRS ones and the MRS method improves the outage performance of cooperative communication system more effectively under outdated CSI than under non‐outdated CSI.

With the rapid popularity of autonomous aerial vehicles (AAV)-Internet of Things (IoT) networks, timely and energy-efficient data collection is a critical issue. The strong Line-of-Sight (LoS) communication of UAVs makes them more susceptible to jamming attacks, resulting in increased communication latency and energy consumption (EC). In this article, the problem of jointly minimizing Age of Information (AoI) and EC in data collection under malicious jamming attacks is formulated by optimizing the diagonal phase shift matrix of the intelligent reflective surface (IRS), UAV trajectory, and the transmit powers of IoT devices. The formulated problem is solved by an alternating optimization (AO) scheme. Specifically, we first obtain the closed-form solution of the IRS diagonal phase shift matrix by using the quantitative passive beamforming method. Then, the UAV trajectory is optimized by the variable parameter particle swarm annealing (VPPSA) algorithm, and the suboptimal solution of the transmit powers of IoT devices is obtained by the successive convex approximation (SCA) algorithm. Extensive simulation results demonstrate that the proposed algorithm outperforms the benchmark schemes in terms of AoI and EC.