Optimizing short-term load forecasting performance is a challenge due to the non-linearity and randomness of electrical load, as well as the variability of system operating patterns. Existing methods often fail to consider how to effectively combine their complementary advantages and fail to fully capture the internal information in the load sequence, leading to a decrease in accuracy. To achieve accurate and efficient short-term load forecasting, this study proposes a novel power grid load forecasting model that integrates Convolutional Neural Networks (CNNs), Long Short-Term Memory Networks (LSTMs), Multi-Head Self-Attention Mechanism (MHSA), Global Attention Mechanism (GAM), and Channel Attention Mechanism (CAM) to achieve efficient and precise short-term load forecasting. This model aims to address the issue in traditional methods where complex temporal features and important information in power grid load data are not fully captured. Firstly, the CNN module is used to extract high-dimensional spatial features from the load data, and a pooling layer is applied to reduce dimensionality while retaining key information. Then, the Multi-Head Self-Attention mechanism is employed to model the long-range dependencies of the sequence data, enhancing the ability to extract temporal features. Next, the LSTM layer further captures the time dependencies in the load sequence. Subsequently, the Global Attention mechanism helps the model focus more on the most relevant parts of the input sequence, improving the model’s performance and generalization ability. The Channel Attention module is then applied to weight different feature channels, highlighting important information and reducing redundancy. Finally, the flattened output layer produces the forecast results. Experimental validation shows that the proposed CNN-MHSA-LSTM-GAM-CAM model outperforms existing mainstream methods in terms of load forecasting accuracy, providing effective support for the optimized scheduling of smart grids.

Stock price prediction and portfolio optimization are critical research areas in financial markets, as they directly impact investment strategies and risk management. Traditional statistical methods and machine learning approaches have been widely applied to these tasks, but they often fail to fully capture the complex dynamics of financial markets. Traditional statistical methods typically rely on unrealistic assumptions or oversimplified models, neglecting the nonlinear and high-dimensional characteristics of market data. Additionally, deep learning methods, especially temporal convolution networks and graph attention networks, have been introduced in this area and have achieved significant improvements in both stock price prediction and portfolio optimization. Therefore, this study proposes a Spatial–Temporal Graph Attention Network (STGAT) that integrates STL decomposition components and graph structures to model both temporal patterns and asset correlations. By combining graph attention mechanisms with temporal convolutional modules, STGAT effectively processes spatiotemporal data, enhancing the accuracy of stock price predictions. Empirical experiments on the CSI 500 and S&P 500 datasets demonstrate that STGAT outperforms other deep learning models in both prediction accuracy and portfolio performance. The investment portfolios constructed based on STGAT’s predictions achieve higher returns in real market scenarios, which validates the feasibility of spatiotemporal feature fusion for stock price prediction and highlights the advantages of graph attention networks in capturing complex market characteristics. This study not only provides a robust tool for portfolio optimization but also offers valuable insights for future research in intelligent financial systems.

Unit commitment (UC) is a critical component for the power system dispatching departments. Current methodologies for solving UC problems predominantly rely on mixed-integer linear programming and are supplemented by data-driven approaches. These methodologies have two primary limitations: first, as the scale of the power grid expands, the complexity of algorithms increases sharply. Second, they fail to fully exploit grid topology information and historical data trends. To address these limitations, this article proposes a constrained graph-temporal convolutional network, which addresses the UC problem by directly predicting power output with constraints. The algorithm takes historical load data from grid nodes as input, utilizes graph convolutional networks to capture the physical grid topology information, and employs temporal convolutional networks to extract temporal features. Ultimately, the output of the graph-temporal convolutional network is projected into the feasible domain through a linear constraint activation layer, achieving accurate power prediction. Experiments conducted on the IEEE 30-BUS and IEEE 118-BUS systems validate the feasibility and superiority of our method in terms of accuracy and computational efficiency.

High-precision interpretable forecasting of multi-energy loads is crucial for planning and optimizing Integrated Energy Systems (IES). Current methods have three key limitations: 1) reliance on closed-box models obscures feature interactions; 2) traditional modal decomposition methods ignore cross-modal correlations and fail to efficiently extract global cycle-trend characteristics; 3) limitations in sampling frequency prevent capturing fast dynamic characteristics of energy loads. To address these, we propose a dual resolution channel multi-period cross reconstruction parallel closed-form continuous-time network (DMP-PCFC) for multi-energy load forecasting. This method includes a multi-period cross-reconstruction network which equivalent to the signal sifter for decoupling original sequences and nonlinear mapping across periods, and models hidden states of energy loads using neuron state iterative differential equations. A parallel Closed-form continuous-time neural network (PCFC) based on chunked parallelization reduces inference time from $O(T)$ (change with input sequence length) to $O(P)$ (fixed cycle length). To overcome inherent sampling resolution limitations, a dual-resolution channel model with weight-sharing PCFC is used for simultaneous inference. Extensive experiments and continuous wavelet transform (CWT) on the IES dataset at Arizona State University’s Tempe campus show that DMP-PCFC offers higher prediction accuracy and greater interpretability than existing methods, providing a new paradigm for downstream tasks based on other research methods, making it valuable for IES engineering applications. The code is publicly available at https://github.com/nuist-xf/DMP-PCFC

Accurate and efficient multienergy load forecasting is fundamental for the secure and economic operation of integrated energy systems (IESs). However, current methods face several critical challenges as follows. 1) Insufficient perception of low-energy transient features leads to significant deviations in marginal scenario predictions. 2) Existing architectures struggle to simultaneously capture long-term trends and short-term dynamics efficiently. 3) Many high-accuracy models suffer from high computational complexity, limiting their practical deployment. To overcome these limitations, we propose the frequency-enhanced seasonal-trend recurrent neural network (FESTRNN), which incorporates two newly designed modules: the energy-enhanced spectral module to amplify critical frequency signals, and the dual-stream collaborative temporal forecasting module to efficiently model multiscale temporal dependencies. Experimental results on the ASU dataset demonstrate that FESTRNN achieves state-of-the-art accuracy across various forecasting horizons with significantly lower computational costs. This study provides a reliable basis for real-time scheduling in IES, offering strong practical value.

In the field of machine learning, continuous-time neural networks (CTNNs) have garnered significant attention due to their powerful modeling capabilities for time series data. However, traditional numerical ordinary differential equation (ODE) solving methods suffer from computational inefficiency. Although closed-form CTNNs (CfCs) alleviate this issue through closed-form approximate solutions, their zero-order approximation method based on segmented constant inputs still exhibits errors when processing high-resolution or irregularly sampled data. This study proposes a numerical integration approximation theory based on the Lagrange interpolation method (NIALIM), which introduces temporal interval factors to linearly approximate the nonlinear integral function in liquid time-constant networks (LTCs) network state equations, deriving closed-form solutions dependent on external inputs and sampling intervals, while rigorously proving the error upper bound between the approximate and original solutions. Furthermore, we construct two novel network architectures: the dynamic feature accumulation closed-form network (DFA-CfN) for general sequence tasks, and the prereconstructed DFA-CfN (PRDFA-CfN) optimized for decomposable sequence tasks, along with six variant networks from these two families. Extensive experiments across six distinct tasks—physical dynamics modeling, human activity recognition, PhysioNet challenge, internet movie database (IMDb) sentiment analysis, XCO₂ remote sensing sequence prediction, and multienergy load forecasting in integrated energy systems (IESs)—demonstrate that our proposed models achieve state-of-the-art performance across all benchmarks, with the highest metrics improvement reaching 69.2%, providing a novel high-performance tool for continuous-time modeling tasks. The code used to support the findings of this study is available from https://github.com/nuist-xf/DFA-CfNs