{"title":"Compact convolution transformer with cross-feature aggregation for hand-gesture recognition","authors":"Satya Narayan , Praful Hambarde , Santosh Kumar Vipparthi , Arka Prokash Mazumdar , Subrahmanyam Murala","doi":"10.1016/j.compeleceng.2025.110727","DOIUrl":"10.1016/j.compeleceng.2025.110727","url":null,"abstract":"<div><div>Hand Gesture Recognition (HGR) plays a crucial role in intuitive human–computer interaction but continues to face challenges such as complex backgrounds, lighting variations, occlusions, and limited training data. To overcome these issues, we propose a Cross Feature Aggregation Compact Convolution Transformer (CrFe-CCT) that integrates multiscale convolutional features with a lightweight transformer architecture. In the proposed CrFe-CCT network, includes the multi-scale Cross Feature Aggregation (CrFe) and CCT modules. The CrFe module help to enhances feature robustness by fusing contextual information across scales, leading to improved recognition accuracy while maintaining low computational complexity. Also, CCT module help to preserve local spatial relationships. Unlike conventional transformers that rely on large-scale data, CrFe-CCT enables efficient learning on both small and large datasets. The experimental results demonstrate that the proposed CrFe-CCT outperforms existing state-of-the-art approaches on subject-dependent datasets, achieving accuracies of 91.95%(HGR-1), 97.70% (MUGD Set1), 95.50% (MUGD Set2), 99.06%(MUGD Set3), 99.82% (NUS-II), 99.90% (ASL-Finger Spelling (FS), and 96.80% (OUHands). On subject-independent datasets, the CrFe-CCT network achieves 40.43%(HGR-1), 85.11% (MUGD), 70.34 (NUS-II Dataset), 82.20% (ASL-Finger Spelling (FS)), respectively. Furthermore, it demonstrates superior efficiency with parameters, memory usage, FLOPs, inference time, and a throughput of images for real-world HGR applications.</div><div>The source code is available at <span><span>https://github.com/satyantazi/CrFe-CCT</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110727"},"PeriodicalIF":4.9,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145158480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Soufiane Ghafiri , Dhaker Abbes , João Pedro F. Trovão , Arnaud Davigny , Maxime Darnon
{"title":"Multi-objective optimization of nanogrids for remote telecom base stations in Canada","authors":"Soufiane Ghafiri , Dhaker Abbes , João Pedro F. Trovão , Arnaud Davigny , Maxime Darnon","doi":"10.1016/j.compeleceng.2025.110714","DOIUrl":"10.1016/j.compeleceng.2025.110714","url":null,"abstract":"<div><div>The telecommunications sector targets net-zero emissions by 2050, yet many remote Canadian base stations rely on diesel generators, incurring high costs and emissions. Most hybrid renewable energy system (HRES) studies overlook snow accumulation, limiting relevance in northern climates. This work proposes a snow-aware hybrid nanogrid for a telecom base station in Dorval Lodge, Quebec, using bifacial PV modules, lithium iron phosphate (LFP) batteries, and a diesel generator. A preliminary HOMER Pro study showed 99% renewable penetration is technically possible but at high cost and without snow, bifacial, or aging effects. We developed a high-fidelity model including hourly snow coverage, seasonal albedo, battery aging, and diesel fuel emission behavior. A joint multi-objective optimization minimizing life cycle cost (LCC) and annual <span><math><msub><mrow><mi>CO</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> under <span><math><mrow><mi>L</mi><mi>P</mi><mi>S</mi><mi>P</mi><mo><</mo><mn>0</mn><mo>.</mo><mn>0001</mn><mtext>%</mtext></mrow></math></span> was solved using a Controlled Elitist NSGA-II algorithm. Three stages were tested: baseline, fixed controls, and monthly adaptive controls. The adaptive strategy achieved the largest gains, cutting <span><math><msub><mrow><mi>CO</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> by 18.59% and LCC by 5.26% versus baseline, with the most sustainable setup using 856 L/year (2.93 t <span><math><msub><mrow><mi>CO</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span>). Sensitivity analysis showed snow-aware designs avoid up to 40.9% higher LCC and 139.7% more <span><math><msub><mrow><mi>CO</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span> seen in snow-unaware cases. Integrating climate-specific snow modeling with adaptive controls enhances economic and environmental performance, offering a robust, transferable solution for remote telecom power in harsh climates.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110714"},"PeriodicalIF":4.9,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145158481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Impact analysis of renewable energy resources and electric vehicles in hybrid power systems","authors":"Anil Kumar, Saurabh Chanana, Amit Kumar","doi":"10.1016/j.compeleceng.2025.110729","DOIUrl":"10.1016/j.compeleceng.2025.110729","url":null,"abstract":"<div><div>This study concentrates on improving load frequency control (LFC) methods for integrated power networks, particularly addressing the fluctuating attributes of energy from renewable sources and electric vehicles. A modified fractional order controller (i.e., fractional order integral-proportional derivative with filter (FOI-PDF)) has been built for the system being studied. Additionally, a new optimization named the Electric Eel Foraging Optimisation (EEFO) has been introduced for improving the settings of various controller parameters. The proposed system under analysis is mathematically modelled and examined to include hydro power plants (HPPs), thermal power plants (TPPs), and gas power plants (GPPs) in each of the two interconnected hybrid power systems. Furthermore, to accommodate case studies, both control areas connect intermittent power from wind power plants (WPPs) & solar power plants (SPPs) along with electric vehicles (EVs) and also examine the effect of communication time (CT) delay. The proposed EEFO optimisation technique surpasses earlier meta-heuristic optimization techniques (MOTs) like (Whale Optimisation Algorithm (WOA), Sine Cosine Algorithm (SCA), Quadratic Interpolation Optimisation (QIV), Arithmetic Optimisation Algorithm (AOA), and Ant Lion Optimisation (ALO)) in terms of convergence curve and the objective function of integral time absolute error (ITAE) value. The ITAE value of EEFO is 88.74%, 88.99%, 5.54%, 90.51%, and 90.27% lower than the values of WOA, SCA, QIV, AOA, and ALO, respectively. A thorough evaluation of several scenarios, including step, multistep, and random disturbances, has been carried out to assess the effectiveness of the suggested control method in contrast to current controllers. In the case of step load disturbances (SLDs), the settling time of the EEFO-based FOI-PDF is 46.05% faster than the recently developed fractional order integral derivative-tilt (FID-T) controller in ΔF<sub>1</sub>, 19.65% faster in ΔF<sub>2</sub>, and 63% faster in ΔP<sub>tie</sub>, respectively. The comprehensive data investigations indicate that the anticipated hybrid power system is the subject of a dynamic performance study that is both superior and enhanced. Additionally, the stability study, encompassing Bode plots and eigenvalues along with sensitivity analysis, has been conducted. The proposed methodology has been validated by an empirical inquiry carried out in real real-time simulator using the OPAL-RT platform.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110729"},"PeriodicalIF":4.9,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145158483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Hilbert transformed differential evolution based optimization of convolution neural network for Electroencephalogram signal filtering and classification","authors":"Raja Sekhar Banovoth, Kadambari K V","doi":"10.1016/j.compeleceng.2025.110726","DOIUrl":"10.1016/j.compeleceng.2025.110726","url":null,"abstract":"<div><div>Electroencephalography (EEG) is widely regarded as an effective non-invasive technique for measuring neuronal activity in the human brain, offering high temporal resolution. However, EEG signals are often contaminated by noise and artifacts, which significantly impact their analysis. To tackle this challenge, various advancements in deep learning-based denoising techniques have been developed. Despite these advancements, identifying the optimal network architecture across a broad range of initial parameters remain complex. Manually tuning these parameters to achieve optimal performance is time-intensive and demands substantial expertise. Along the same lines, we introduce a novel residual block-based neural network for automatic artifact removal and a convolution neural network for classification. The architecture unfolds in two distinct stages: Firstly, the noise elimination, where the raw EEG signals undergo the Hilbert transformation, converting them into complex-valued signals. These signals serve as the input for a Residual block-based Convolution Neural Network constructed using the innovative technique of Differential Evolution (HT-DEResNet), which optimizes both the architecture and initial parameters. This enables the effective removal of artifacts without manual intervention. The model is applied across three complex datasets: HaLT, eye artifact, and major depressive disorder. The secondary stage comprises the development of a Multi-Convolution Neural Network (MCNN) for classification. To validate the effectiveness, the BCI competition datasets 2a & 2b are subjected to the HT-DEResNet framework. Subsequently, the signals are classified. The results not only surpass state-of-the-art methods in accuracy but also achieve a 17.73% improvement in performance attributed to noise removal.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110726"},"PeriodicalIF":4.9,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145158482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"SecureNet: A deep learning inspired security framework for healthcare data","authors":"Vishnu Bharadwaj Bayari Parkala , Gaurav Bhatnagar , Chiranjoy Chattopadhyay","doi":"10.1016/j.compeleceng.2025.110723","DOIUrl":"10.1016/j.compeleceng.2025.110723","url":null,"abstract":"<div><div>As medical devices become increasingly interconnected through the Internet of Medical Things (IoMT), safeguarding image data against unauthorized access and tampering has become a pressing challenge. Many existing solutions fall short in balancing computational efficiency with robust encryption, particularly when high-fidelity recovery of diagnostic images is required. This work presents SecureNet, a hybrid encryption framework tailored for medical imaging applications. This work proposes SecureNet, a compact and resilient encryption framework designed for secure medical image transmission. SecureNet leverages a convolutional autoencoder to extract compact latent features, applies spatial rearrangement using a Hilbert curve to disrupt pixel locality, and diffusion with a chaos-driven random projection. The result is a highly randomized and distorted representation, complicating any attempt at unauthorized analysis. The encrypted data can be accurately recovered through symmetric decryption operations. Experimental results, supported by security and comparative analyses, validate the effectiveness and generalizability of the proposed framework in resisting various attacks, demonstrating its encryption efficacy and performance on par with state-of-the-art approaches. The framework presents a scalable solution for securing healthcare data in dynamic and resource-constrained IoMT environments.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110723"},"PeriodicalIF":4.9,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145128274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A low-power switching technique for SAR ADCs achieving energy efficient transitions at the input and reference voltage sources","authors":"Mehdi Sotoudeh, Mehdi Habibi","doi":"10.1016/j.compeleceng.2025.110728","DOIUrl":"10.1016/j.compeleceng.2025.110728","url":null,"abstract":"<div><div>This study presents an alternative approach to capacitor switching aimed at lowering switching energy in successive approximation register analog-to-digital converters (SAR ADCs). This approach eliminates energy consumption from the input voltage by establishing a direct connection to the comparator’s positive terminal. To mitigate offset drawbacks, a low-power offset-cancellation comparator is employed. Furthermore, by isolating all capacitors connected to the comparator’s negative terminal during the switching of reference voltages, the energy overhead introduced by toggling the reference voltages is also ideally reduced to zero. This achieves a theoretical 100% reduction in switching energy compared to conventional SAR ADC methods under idealized assumptions. In practice, non-ideal factors such as parasitic capacitance, charge injection, comparator dynamic power, and control logic overhead contribute to residual energy consumption. This energy reduction is achieved while maintaining a low complexity in the control block. Additionally, this architecture employs only 25% of the capacitors required in traditional design, leading to substantial hardware savings. The suggested 8-bit SAR ADC was simulated in a 65 nm CMOS process with a 1.2 V power supply, 0.9 V reference voltage, and an operating sampling rate of 1 MS/s. As observed in the simulation data, the effective number of bits, power dissipation, and figure of merit are determined to be 6.9 bits, 1.91 µW, and 15.99 fJ/conversion-step, respectively. To evaluate the converter's efficiency, Monte Carlo, corner, and temperature simulations were conducted, demonstrating satisfactory performance.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110728"},"PeriodicalIF":4.9,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145128255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hassan Harb , Fouad Al Tfaily , Kassem Danach , Hussein Hazimeh , Ali Jaber
{"title":"MOTION: Multi-models correlation framework for energy-saving in wireless video sensor networks","authors":"Hassan Harb , Fouad Al Tfaily , Kassem Danach , Hussein Hazimeh , Ali Jaber","doi":"10.1016/j.compeleceng.2025.110720","DOIUrl":"10.1016/j.compeleceng.2025.110720","url":null,"abstract":"<div><div>Wireless Video Sensor Networks (WVSNs) face critical challenges in energy consumption and data bandwidth utilization due to multiple sensor nodes transmitting redundant data of the observed phenomenon. Thus, identifying and reducing such redundancies is becoming essential for extending the network lifetime and emphasizing the quality of the collected data. In this paper, we propose Multi-mOdels correlaTION framework (MOTION) that efficiently removes data duplication and conserve sensor energies in WVSNs. MOTION proposes new data reduction methods based on the temporal–spatial correlations that could be applied at different node levels, e.g. sensors and cluster-heads (CHs). At the sensor level, MOTION introduces two temporal-based correlation mechanisms to search the similarity among frames collected during each period; the first mechanism aims to detect short-term duplication, e.g. among consecutive frames, while the second one allows to detect scene changes then trigger transmissions when the variation is significant. At the second level, CH geographically groups video sensors into clusters to find the spatial correlation between them, then a scheduling strategy is applied to switch spatially-correlated ones into sleep/active modes. Extensive simulations using real-world video data sets as a benchmark are conducted to show the efficiency of the proposed framework. The results demonstrated that MOTION can reduce up to 92.4% of collected video data, leading to tremendous energy savings and network lifetime up to 74.3% compared to existing approaches.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110720"},"PeriodicalIF":4.9,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145128273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Revolutionizing plant disease diagnosis through vision-based intelligence and next-generation computing","authors":"Amreen Batool , Yung-Cheol Byun","doi":"10.1016/j.compeleceng.2025.110695","DOIUrl":"10.1016/j.compeleceng.2025.110695","url":null,"abstract":"<div><div>Plant diseases significantly threaten global agriculture by reducing crop quality and yield. Early and accurate detection is vital to mitigate these impacts and ensure food security. This review presents a comprehensive survey of vision-based machine learning (ML) and deep learning (DL) approaches for plant disease detection. This review introduces a comparative analysis of more than 25 benchmark datasets and categorizes the progression from traditional ML methods to advanced DL models such as CNN, GAN, and Vision Transformers. This paper also addresses practical challenges such as image noise, environmental variability, and the domain gap between controlled and real-world datasets. Furthermore, the review explores the integration of Large Language Models (LLMs) into plant disease monitoring pipelines for annotation assistance, real-time farmer interaction, and multimodal reasoning. Moreover, the study emphasizes mobile and edge AI applications, including smartphone-based tools, AR interfaces, and IoT-enabled monitoring, enhancing accessibility for farmers in resource-constrained environments. A novel gap analysis and research roadmap are proposed to differentiate from existing works, outlining a future AI-driven agricultural ecosystem. This review concludes by identifying critical challenges and offering actionable research directions for robust, scalable, and interpretable plant disease detection systems in real-world agricultural settings.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110695"},"PeriodicalIF":4.9,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145118370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mirza Ateeq Ahmed Baig , Naeem Iqbal Ratyal , Adil Amin , Umar Jamil , Haris M. Khalid , Muhammad Fahad Zia
{"title":"A multi-modal deep learning framework for power quality disturbance classification: An integration of 1D time-series signals and 2D scalograms","authors":"Mirza Ateeq Ahmed Baig , Naeem Iqbal Ratyal , Adil Amin , Umar Jamil , Haris M. Khalid , Muhammad Fahad Zia","doi":"10.1016/j.compeleceng.2025.110716","DOIUrl":"10.1016/j.compeleceng.2025.110716","url":null,"abstract":"<div><div>Power quality (PQ) is crucial for the dependable functioning of electrical systems, requiring stable voltage, frequency, and waveform integrity. However, power quality disturbances (PQDs), resulting from faults, nonlinear loads, and switching events, can degrade system performance, damage equipment, and reduce operational efficiency. Accurate identification and classification of PQDs are therefore critical for effective mitigation. Traditional methods that rely solely on either one-dimensional (1D) time-series signals or two-dimensional (2D) waveform images often fail to capture the full characteristics of disturbances, leading to reduced accuracy. To address this limitation, a multi-modal deep learning framework is proposed that integrates 1D time-series data with corresponding 2D scalogram images. The proposed model employs parallel 1D and 2D convolutional neural networks (CNNs), each enhanced with attention mechanisms to enhance feature extraction by focusing on modality-specific salient information. The proposed model is evaluated on a comprehensive synthetic dataset of sixteen PQD types. Experimental results demonstrate that the proposed approach achieves an average classification accuracy of 99.99%, a sensitivity of 99.98%, and a specificity of 99.99%, outperforming existing methods. These results demonstrate the framework’s robustness and its potential as an effective solution for PQD monitoring and classification in smart grid environments.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110716"},"PeriodicalIF":4.9,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145118369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"SAGANConvLSTM: A novel spatio-temporal forecasting approach combining semivariogram-enhanced GAN and ConvLSTM for power load forecasting","authors":"Rasoul Jalalifar , Mahmoud Reza Delavar , Seyed Farid Ghaderi , Seyedeh Leyla Mansouri Alehashem","doi":"10.1016/j.compeleceng.2025.110718","DOIUrl":"10.1016/j.compeleceng.2025.110718","url":null,"abstract":"<div><div>In power distribution networks, spatio-temporal load forecasting plays a crucial role in decision-making and the development of distribution networks. Accurate forecasting models are essential to handle the complex dependencies in power consumption across different districts of megacities. However, the presence of missing values in smart meter data often caused by device malfunctions or communication failures, can significantly degrade model performance by increasing complexity and reducing forecasting accuracy. To address this challenge, this paper presents a novel forecasting approach named SAGAN<img>ConvLSTM based on Spatial Autocorrelation (SA) utilizing spatio-temporal semivarigram and Generative Adversarial Network (GAN) to combine spatial statistics and deep learning to impute missing values in power load time series. The semivariogram quantifies spatial and temporal dependencies among substations and guides the GAN to reconstruct missing data while preserving realistic spatio-temporal structures. After imputation, the refined time series is processed by a Convolutional Long Short-Term Memory (ConvLSTM) network to generate short-term power load forecasts.The proposed model was evaluated using data from Tehran's power distribution network, achieving a mean absolute error (MAE) of 10.05 % and root mean square error (RMSE) of 15.46 %, outperforming other models like GRU, LSTM, ConvLSTM, and SA-ConvLSTM in forecasting power load over a 20-day period.</div></div>","PeriodicalId":50630,"journal":{"name":"Computers & Electrical Engineering","volume":"128 ","pages":"Article 110718"},"PeriodicalIF":4.9,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145118368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}