Average Error Reduction Research Articles

Efficient and Secure Traffic Scheduling Based on Private Sketch

In today’s data–driven world, the explosive growth of network traffic often leads to network congestion, which seriously affects service performance and user experience. Network traffic scheduling is one of the key technologies to deal with congestion problems. Traditional traffic scheduling methods often rely on static rules or pre–defined policies, which make it difficult to cope with dynamically changing network traffic patterns. Additionally, the inability to efficiently manage tail contributors that disproportionately contribute to traffic can further exacerbate congestion issues. In this paper, we propose ESTS, an efficient and secure traffic scheduling based on private sketch, capable of identifying tail contributors to adjust routing and prevent congestion. The key idea is to develop a randomized admission (RA) structure, linking two count–mean–min (CMM) sketches. The first CMM sketch records cold items, while the second, following the RA structure, stores hot items with high frequency. Moreover, considering that tail contributors may leak private information, we incorporate Gaussian noise uniformly into the CMM sketch and RA structure. Experimental evaluations on real and synthetic datasets demonstrate that ESTS significantly improves the accuracy of feature distribution estimation and privacy preservation. Compared to baseline methods, the ESTS framework achieves a 25% reduction in average relative error and a 30% improvement in tail contributor identification accuracy. These results underline the framework’s efficiency and reliability.

Investigating the accuracy of adjusting for examiner differences in multi-centre Objective Structured Clinical Exams (OSCEs). A simulation study of video-based Examiner Score Comparison and Adjustment (VESCA)

IntroductionEnsuring examiner equivalence across distributed assessment locations is a priority within distributed Objective Structured Clinical Exams (OSCEs) but is challenging as examiners are typically fully nested within locations (i.e. no overlap in performances seen by different groups of examiners). Video-based Examiner Score Comparison and Adjustment (VESCA) is a recently developed method which uses video-based linking to compare and (potentially) adjust for the effect of different groups of examiners within OSCEs. Whilst initial research on VESCA has been promising, the accuracy of the resulting adjusted scores is unknown. Given this, we aimed to investigate the accuracy of adjusted scores produced by VESCA under a range of plausible operational parameters.MethodsUsing statistical simulation, we investigated how: 1/proportion of participating examiners, 2/ number of linking videos, 3/baseline differences in examiner stringency between schools (i.e. whether examiners in School A are, on average, more stringent than the examiners in School B), 4/number of OSCE stations and 5/different degrees of random error within examiners’ judgements influenced accuracy of adjusted scores.We generated distributions of students’ “true” performances across several stations, added examiner error, and simulated linking through crossed video-scoring (as occurs in VESCA). We then used Many Facet Rasch Modelling to produce an adjusted score for each student which we compared with their corresponding original “true” performance score. We replicated this 1000 times for each permutation to determine average error reduction and the proportion of students whose scores became more accurate. Simulation parameters were derived from a real, summative, whole curriculum undergraduate Year 3 OSCE at Keele University School of Medicine.ResultsWe found that in all conditions where no baseline difference existed between groups of examiners, score adjustment only minimally improved or even worsened score accuracy. Conversely, as the size of baseline differences between schools increased, adjustment accuracy increased, reducing error by up to 71% and making scores more accurate for up to 93% of students in the 20% baseline-difference condition.ConclusionsScore adjustment through VESCA has the potential to substantially enhance equivalence for candidates in distributed OSCEs in some circumstances, whilst making scores less accurate in others. These findings will support judgements about when score adjustment may beneficially aid OSCE equivalence.

A new modeling approach for microplastic drag and settling velocity

Understanding microplastics' (MPs') transport and settling behaviors in aquatic environments is crucial for devising effective management strategies. This study contributes a novel modeling framework to develop accurate and interpretable drag and velocity models for MPs using machine learning techniques. It achieves faster model creation and improved accuracy than traditional methods like theoretical analysis and data fitting. The framework demonstrates high predictive accuracy across different MP types (1D, 2D, 3D, and mixed), with a coefficient of determination CD = 0.86–0.95 for the drag models and CD = 0.92–0.95 for the velocity models. Compared with best-performing empirical approaches, the new drag models exhibit an average reduction in root mean square error (RMSE) by 59% and mean absolute error (MAE) by 62%. Similarly, the velocity models show a mean decrease in RMSE and MAE by 27% and 25%, respectively. Moreover, the framework outperforms commonly used symbolic regression methods, reducing errors by 18%–27%. The sensitivity analysis reveals that the relative density difference and the dimensionless diameter are essential for predicting the settling of all MP types, while the effective shape parameters vary across different MP categories. By providing accurate predictions of MPs' settling dynamics, this study offers insights for developing targeted mitigation strategies to reduce MPs' environmental impacts.

Drone Photogrammetry-based Wind Field Simulation for Climate Adaptation in Urban Environments

Addressing climate change issues is one of the most important tasks within the United Nations Sustainable Development Goals. Accurate and efficient simulation of wind fields within cities is essential for climate adaptation. Traditional simplified geometric model-based wind flow simulation can lead to significant errors, affecting the ability to develop effective urban climate strategies. This study addresses this limitation by introducing a novel workflow that leverages drone photogrammetry, deep learning, and geometric complexity quantification to create highly detailed 3D models of in-use building clusters within cities. These models are subsequently used for computational fluid dynamics simulations to accurately predict urban wind fields. The proposed method was validated on three real-world building clusters. Compared to traditional footprint extrusion models, the proposed method demonstrates an average error reduction of 29.2% in large eddy simulation cases and 17.6% in steady Reynolds-averaged Navier-Stokes equations cases. Meanwhile, the proposed model improved computational efficiency by an average of 33.7% in large eddy simulations compared to the flashy oblique photography model. The proposed method provides a balanced model of accuracy and efficiency for urban flow simulations. It has the potential to be incorporated into computational fluid dynamics best practice guidelines, thereby promoting the development of climate-resilient cities.

ACT‐GAN: Radio map construction based on generative adversarial networks with ACT blocks

AbstractThe radio map serves as a vital tool in assessing wireless communication networks and monitoring radio coverage, providing a visual representation of electromagnetic spatial characteristics. To address the limitation of low accuracy in current radio map construction method, this article presents a novel method based on Generative Adversarial Network (GAN), called ACT‐GAN. This method incorporates the aggregated contextual‐transformation block, the convolutional block attention module, and the transposed convolutional block into the generator, significantly enhancing the construction accuracy of radio map. The performance of ACT‐GAN is validated in three distinct scenarios. The results indicate that, in scenario 1, where the transmitter locations are known, the average reduction in Root Mean Square Error (RMSE) is 14.6%. In scenario 2, where the transmitter locations are known and supplemented with sparse measurement maps, the average reduction in RMSE is 13.3%. Finally, in scenario 3, where the transmitter locations are unknown, the average reduction in RMSE is 7.1%. Moreover, the proposed model exhibits clearer predictive results and can accurately capture multi‐scale shadow fading.

PB-Trajectron: Physics bounded neural network for generalized trajectory prediction

Vehicle trajectory prediction is a critical technology in autonomous driving systems, as the quality of prediction directly affects downstream path planning and vehicle control. Although recent studies have shown that models combining kinematic rules with data-driven methods exhibit better interpretability in trajectory prediction, these models often adopt fixed constraint strengths, limiting their generalization ability in diverse traffic scenarios. This fixed constraint strength design restricts the model’s adaptability to different traffic environments.To address this issue, we propose PB-Trajectron, a dynamic integration mechanism that enhances the model’s prediction performance and generalization capability. PB-Trajectron is a dynamic integration mechanism that combines vehicle kinematic rules with deep learning prediction models for generalized vehicle trajectory prediction. The model integrates physics-based motion models and Deep Neural Networks (DNNs) to provide reasonable physical explanations for predicting vehicle trajectories at different speeds. First, we establish two vehicle kinematic constraints applicable to different prediction scenarios, enabling the agent to switch between these constraints based on the causal relationships between vehicle motion states to avoid generating non-physical trajectory results. Second, we propose a kinematic constraint decision framework based on velocity thresholds, which demonstrates adaptive adjustment, allowing the agent to switch tasks according to real-time state conditions and adaptively adjust the contribution of different physical constraints based on actual vehicle speeds. Finally, we investigate the advantages of PB-Trajectron in Out-of-Domain(OOD) generalization.Cross-scenario experiments on the nuScenes dataset show that, compared to previous methods, PB-Trajectron reduces the final displacement error by 7.69% when the prediction step length is 4. Furthermore, out-of-domain generalization tests on the INTERACTION dataset demonstrate that PB-Trajectron achieves a 12.23% reduction in average prediction error on datasets from different countries and scenarios compared to previous methods. The proposed mechanism can better adapt to complex and diverse traffic scenarios, laying the foundation for explainable and robust autonomous driving systems.

Local Weather and Global Climate Data-Driven Long-Term Runoff Forecasting Based on Local–Global–Temporal Attention Mechanisms and Graph Attention Networks

The accuracy of long-term runoff models can be increased through the input of local weather variables and global climate indices. However, existing methods do not effectively extract important information from complex input factors across various temporal and spatial dimensions, thereby contributing to inaccurate predictions of long-term runoff. In this study, local–global–temporal attention mechanisms (LGTA) were proposed for capturing crucial information on global climate indices on monthly, annual, and interannual time scales. The graph attention network (GAT) was employed to extract geographical topological information of meteorological stations, based on remotely sensed elevation data. A long-term runoff prediction model was established based on long-short-term memory (LSTM) integrated with GAT and LGTA, referred to as GAT–LGTA–LSTM. The proposed model was compared to five comparative models (LGTA–LSTM, GAT–GTA–LSTM, GTA–LSTM, GAT–GA–LSTM, GA–LSTM). The models were applied to forecast the long-term runoff at Luning and Pingshan stations in China. The results indicated that the GAT–LGTA–LSTM model demonstrated the best forecasting performance among the comparative models. The Nash–Sutcliffe Efficiency (NSE) of GAT–LGTA–LSTM at the Luning and Pingshan stations reached 0.87 and 0.89, respectively. Compared to the GA–LSTM benchmark model, the GAT–LGTA–LSTM model demonstrated an average increase in NSE of 0.07, an average increase in Kling–Gupta Efficiency (KGE) of 0.08, and an average reduction in mean absolute percent error (MAPE) of 0.12. The excellent performance of the proposed model is attributed to the following: (1) local attention mechanism assigns a higher weight to key global climate indices at a monthly scale, enhancing the ability of global and temporal attention mechanisms to capture the critical information at annual and interannual scales and (2) the global attention mechanism integrated with GAT effectively extracts crucial temporal and spatial information from precipitation and remotely-sensed elevation data. Furthermore, attention visualization reveals that various global climate indices contribute differently to runoff predictions across distinct months. The global climate indices corresponding to specific seasons or months should be selected to forecast the respective monthly runoff.

A simplified model for seismic risk assessment of three-dimensional irregular steel moment frame buildings

Seismic risk assessment of irregular buildings requires nonlinear dynamic analysis and three-dimensional (3D) models. However, creating and analyzing such models can be intricate and time-consuming, making the assessment process challenging. Furthermore, this complexity escalates when evaluating building stocks or generating a dataset of buildings with diverse attributes for use in learning algorithms. Therefore, this paper proposes an energy-based model for irregular steel moment-resisting frame buildings to simplify 3D models and improve the time efficiency of structural assessment compared to conventional 3D models. The proposed model is derived from the fish-bone model but has enhanced configuration, elastic, and inelastic properties. These improvements aim to broaden its applicability to encompass 3D irregular mid-rise buildings. Four-, six-, and nine-story irregular buildings in plan and height are utilized to validate the proposed model. The responses investigated in these buildings include eigenvalues, pushover curves, time history responses, incremental dynamic analysis results, fragility curves, and risk analysis outcomes. The findings demonstrate that the simplified model responses are in close agreement with those of 3D models. The comparison of results between the proposed simplified model and existing models indicates that the accuracy of the simplified model is at least 2.2 times that of the existing models. Moreover, it provides an average error reduction of 60 % in contrast to the existing models. These advancements have been achieved while the proposed simplified model maintains a level of time efficiency comparable to the current models. The results from the analysis of the created buildings using the proposed simplified model indicate that the model is suitable for application in the seismic risk assessment of building inventories and generating datasets for future research involving machine learning-based methodologies.

Explainable Encoder-Prediction-Reconstruction Framework for the Prediction of Metasurface Absorption Spectra.

The correlation between metasurface structures and their corresponding absorption spectra is inherently complex due to intricate physical interactions. Additionally, the reliance on Maxwell's equations for simulating these relationships leads to extensive computational demands, significantly hindering rapid development in this area. Numerous researchers have employed artificial intelligence (AI) models to predict absorption spectra. However, these models often act as black boxes. Despite training high-performance models, it remains challenging to verify if they are fitting to rational patterns or merely guessing outcomes. To address these challenges, we introduce the Explainable Encoder-Prediction-Reconstruction (EEPR) framework, which separates the prediction process into feature extraction and spectra generation, facilitating a deeper understanding of the physical relationships between metasurface structures and spectra and unveiling the model's operations at the feature level. Our model achieves a 66.23% reduction in average Mean Square Error (MSE), with an MSE of 2.843 × 10-4 compared to the average MSE of 8.421×10-4 for mainstream networks. Additionally, our model operates approximately 500,000 times faster than traditional simulations based on Maxwell's equations, with a time of 3×10-3 seconds per sample, and demonstrates excellent generalization capabilities. By utilizing the EEPR framework, we achieve feature-level explainability and offer insights into the physical properties and their impact on metasurface structures, going beyond the pixel-level explanations provided by existing research. Additionally, we demonstrate the capability to adjust absorption by changing the metasurface at the feature level. These insights potentially empower designers to refine structures and enhance their trust in AI applications.

Numerical model of asphaltene deposition in vertical wellbores: Considerations of particle shape and drag force

Asphaltene deposition and plugging are common phenomena during crude oil production, significantly reducing production efficiency. Understanding the deposition behavior of asphaltene particles in vertical wellbores, particularly in relation to drag force interactions and settlement velocity, is imperative to mitigate these issues. Existing empirical models often fail due to their inherent inaccuracies and neglect of asphaltene particles' irregular shapes. This work established a quantitative method to describe irregular geometries of asphaltene particles. Shape coefficient, ∅, and shape coefficient influence factor, θ, are used to assess the impact of particle shape on the drag force coefficient and settling velocity. Experimental results from sedimentation experiments reveal particles with smaller shape coefficients encounter greater flow resistance and higher drag force coefficients. Moreover, the impact of shape coefficient on the drag coefficient gradually decreases as the Reynolds number increases. Based on these experimental findings, distinct models for the drag force coefficient and settling velocity of asphaltene particles are established, achieving a reduction in average error ranging between 2 % ∼ 25 % compared to existing models.

Timing offset calibration for TOF PET using stationary line source scans at multiple positions

Background. Accurate timing offset calibration is crucial for time-of-flight (TOF) positron emission tomography (PET) to mitigate image artifacts and improve quantitative accuracy. However, existing methods are often time-consuming, complex, or costly.Objective. This paper presents a method for TOF PET timing offset calibration that eliminates the need for costly equipment, phantoms, short-half-life sources, and precise source positioning.Approach. We estimate channel timing offsets using stationary scans of a68Ge line source, typically used for routine quality control, at a minimum of three non-coplanar positions, with each position scanned for two minutes. The line source positions are accurately determined by applying a simple algorithm to their reconstructed images, allowing precise calculation of arrival time differences. Channel timing offsets are estimated by solving a least squares problem. This method is assessed through analyses of phantoms and patient images using a RAYSOLUTION DigitMI 930 scanner.Main results. The estimated timing offsets ranged from -500 ps to 500 ps across all channels. Calibration with a minimum of three scanned positions was sufficient to correct these offsets, achieving less than a 1% discrepancy across various metrics of the image quality (IQ) phantom compared to 12 positions. This calibration significantly reduced edge artifacts in TOF reconstruction of both phantoms and patients. Furthermore, the IQ phantom displayed a 14% increase in average contrast recovery, a 61% reduction in average background variability across all spheres, and a 90% reduction in average residual error. Consistent with the phantom results, patient data revealed enhancements in maximum standardized uptake values (SUVmax) from 14% to 55% for lesions measuring 6 mm to 14 mm. The calibration also improved lesion-to-background contrast and eliminated artifacts caused by the spillover effect of the kidneys and bladder.Significance. The proposed method is fast, user-friendly, and cost-effective, effectively improving lesion detection and diagnostic accuracy.

A novel prediction model of total drag coefficient of an MPV vehicle based on LSA optimization

Abstract Environmental degradation and energy scarcity are caused by the ongoing rise in the number of automobiles. The most efficient way to lower air resistance while driving is to boost vehicle energy efficiency and reduce energy consumption. In this case, a specific MPV model serves as the study subject in this work, and we use CFD simulation to estimate the drag coefficient along with the selection of important factors. Furthermore, to increase prediction accuracy, a lightning search algorithm (LSA)-based BP-neural-network prediction model is created and developed. The results show that the average relative error reduction of 2.83% in the drag resistance model optimized by the lightning search method represents a significant improvement in accuracy. This phenomenon can be explained by the LSA’s ability to resolve the BP neural network’s initial weight and threshold uncertainty as well as the issue of local minima being quickly reached. The high prediction accuracy drag prediction model, obtained herein, is useful for creating and producing automobile shapes with low drag coefficients.

Long-term wind power forecasting with series decomposition and spatio-temporal graph neural network

ABSTRACT The stable operation of the power grid requires accurate predictions of wind power generation and the stabilization of its fluctuations through the integration of other energy sources. An increasing number of deep learning methods are now being employed in the field. However, due to the instability of wind power data, existing methods struggle to uncover deep spatiotemporal dependencies. We propose a novel method named SD-STGNN (Series Decomposition and Spatio-Temporal Graph Neural Network). SD-STGNN first decomposes unstable wind power data into seasonal and trend components. For the seasonal data reflecting short-term fluctuation patterns, we employ a Gated Temporal Convolutional Network and Graph Convolutional Network to capture spatiotemporal relationships. Additionally, for trend data reflecting long-term fluctuation patterns, we introduce a Temporal-Feature Enhancement module, utilizing Multi-Layer Perceptrons to extract deep information along both temporal and feature dimensions. Extensive experiments were conducted on the SDWPF public dataset. Compared to existing state-of-the-art baseline methods, our proposed SD-STGNN model achieves a notable average reduction in Mean Absolute Error by approximately 6.26%, in Root Mean Squared Error by 7.55%, and in Mean Absolute Percentage Error by 2.65%. Additionally, there is an average improvement of about 4.74% in the coefficient of determination.

Developing deep learning surrogate models for digital twins in mineral processing – A case study on data-driven multivariate multistep forecasting

The escalating demand for environmental and social sustainability underscores the critical need for large industries such as mining and metallurgy to function optimally. Achieving optimal operation of a mineral processing plant requires timely access to information and rapid implementation of corrective actions. Essential information, such as current and future production, plays a pivotal role in enhancing process efficiency by facilitating adjustments in key aspects, including feed, control and operating variables. This study delves into the application of deep learning algorithms to forecast the gold concentrate grade of a flotation process half an hour in advance, a valuable tool for short-term decision-making, e.g., with regard to optimising the operation and process control. Multivariate time series data from a real gold processing plant was acquired through the process instrumentation and control system and augmented with simulated data at 5-minute intervals over 236 days. Surrogate models based on LSTM, GRU and CNN architectures were developed and analysed for predicting the gold grade one to six steps ahead. Evaluation of the testing dataset indicates that the CNN exhibits superior performance, achieving an average error reduction of 36 %, 45 % and 35 % for the performance metrics RMSE, MAE and MAPE respectively compared to the baseline model. These findings reveal the high potential of deep learning surrogate methods for accurately forecasting process outputs, enabling effective data-driven process control and optimisation strategies.

Optimisation of real‐scene 3D building models based on straight‐line constraints

AbstractDue to the influence of repeated textures or edge perspective transformations on building facades, building modelling based on unmanned aerial vehicle (UAV) photogrammetry often suffers geometric deformation and distortion when using existing methods or commercial software. To address this issue, a real‐scene three‐dimensional (3D) building model optimisation method based on straight‐line constraints is proposed. First, point clouds generated by unmanned aerial vehicle (UAV) photogrammetry are down‐sampled based on local curvature characteristics, and structural point clouds located at the edges of buildings are extracted. Subsequently, an improved random sample consensus (RANSAC) algorithm, considering distance and angle constraints on lines, known as co‐constrained RANSAC, is applied to further extract point clouds with straight‐line features from the structural point clouds. Finally, point clouds with straight‐line features are optimised and updated using sampled points on the fitted straight lines. Experimental results demonstrate that the proposed method can effectively eliminate redundant 3D points or noise while retaining the fundamental structure of buildings. Compared to popular methods and commercial software, the proposed method significantly enhances the accuracy of building modelling. The average reduction in error is 59.2%, including the optimisation of deviations in the original model's contour projection.

Intelligent RCS Extrapolation Technology of Target Inspired by Physical Mechanism Based on Scattering Center Model

In this paper, a technology named SCM−ANN combining physical scattering mechanisms and artificial intelligence is proposed to realize radar cross-section (RCS) extrapolation of non-cooperative conductor targets with higher efficiency. Firstly, an adaptive scattering center (SC) extraction algorithm is used to construct the scattering center model (SCM) for non-cooperative targets from radar echoes in the low-frequency band (LFB). Secondly, an artificial neural network (ANN) is constructed to capture the nonlinear relationship between the real LFB echoes and those reconstructed from the SCM. Finally, the SCM is used to reconstruct echoes in the high-frequency band (HFB), and these reconstructions, together with the trained ANN, optimize the extrapolated HFB RCS. For the SCM−ANN technology, physical mechanistic modes are used for trend prediction, and artificial intelligence is used for regression optimization based on trend prediction. Simulation results show that the proposed method can achieve a 50% frequency extrapolation range, with an average prediction error reduction of up to 40% compared with the traditional scheme. By incorporating physical mechanisms, this proposed approach offers improved accuracy and an extended extrapolation range compared with the RCS extrapolation techniques relying solely on numerical prediction.

Shear wave anisotropy response characteristics of laminated shale and its correction methods

Laminated shale exhibits strong anisotropic characteristics in horizontal wells, significantly affecting the application of shear wave travel time data. Using equivalent medium theory and the elastic wave equation, we constructed a physical model for laminated shale, simulating the impact of lamination angle and lamination density on shear wave velocity and shear wave anisotropy coefficient. We established the relationship between the shear wave anisotropy coefficient and the cosine of the lamination angle. The simulation results indicate that shear wave velocity decreases with an increase in lamination angle and density, while the shear wave anisotropy coefficient gradually decreases with an increase in lamination angle and initially increases and then decreases with an increase in lamination density. Moreover, it exhibits a power exponential relationship with the cosine of the lamination angle. After correction, the minimum horizontal principal stress and fracture pressure calculated from shear wave data show an average relative error reduction of 6.72% and 7.77%, respectively, compared to measured data. This demonstrates that the correction method effectively eliminates the impact of laminated shale anisotropy on shear waves.

Non-Markovian quantum gate set tomography

Engineering quantum devices requires reliable characterization of the quantum system, including qubits, quantum operations (also known as instruments) and the quantum noise. Recently, quantum gate set tomography (GST) has emerged as a powerful technique for self-consistently describing quantum states, gates, and measurements. However, non-Markovian correlations between the quantum system and environment impact the reliability of GST. To address this, we propose a self-consistent operational framework called instrument set tomography (IST) for non-Markovian GST. Based on the stochastic quantum process, the instrument set describes instruments and system-environment (SE) correlations. We introduce a linear inversion IST (LIST) to describe instruments and SE correlations without physical constraints. The disharmony of linear relationships between instruments is detected. Furthermore, we propose a physically constrained statistical method based on the maximum likelihood estimation for IST (MLE-IST) with adjustable dimensions. MLE-IST shows significant flexibility in adapting to different types of devices, such as noisy intermediate-scale quantum (NISQ) devices, by adjusting the model and constraints. Experimental results demonstrate the effectiveness and necessity of simultaneously describing instruments and SE correlations. Specifically, the LIST and MLE-IST obtains significant improvement on average square error reduction in the imperfect implemented simulations by orders of −23.77 and −6.21, respectively, compared to their comparative methods. Remarkably, real-chip experiments indicate that a polynomial number of parameters with respect to the Markovian order are sufficient to characterize non-Markovian quantum noise in current NISQ devices. Consequently, IST provides an essential and self-consistent framework for characterizing, benchmarking, and developing quantum devices in terms of the instrument set.

Improved Error-Based Ensemble Learning Model for Compressor Performance Parameter Prediction

Large compressors have complex structures and constantly changing operating conditions. It is challenging to build physical models of compressors to analyse their performance parameters. An improved error-based stacked ensemble learning prediction model is proposed in this work. This model simplifies the modelling steps in a data-driven manner and obtains accurate prediction results. An enhanced integrated model employs K-fold cross-validation to assign dataset weights based on validation set errors, achieving a 12.4% reduction in average output error. Additionally, the output error of the meta-model undergoes a Box–Cox transformation for error compensation, decreasing the average output error by 14.0%. The Stacking model, combining the above improvements, notably reduces the root-mean-square errors for power, surge, and blocking boundaries by 24.2%, 20.6%, and 23.3%, respectively. This integration significantly boosts prediction accuracy.

A multi-energy load forecasting method based on complementary ensemble empirical model decomposition and composite evaluation factor reconstruction

Ensuring precise multi-energy load forecasting is crucial for the effective planning, management, and operation of Integrated Energy Systems (IES). This study proposes a novel multivariate load forecasting model based on time-series decomposition and reconstruction to handle the complex, high-dimensional multi-energy load data in IES and enhance forecasting accuracy. Initially, the model conducts a thorough correlation analysis and variable screening to minimize irrelevant data interference. It then applies denoising by decomposing the load sequence into modal components with distinct characteristics, using the complementary ensemble empirical mode decomposition (CEEMD). To overcome the unstable prediction accuracy inherent in time-domain decomposition methods, this study introduces an innovative composite evaluation factor (CEF) that reconstructs the modal components after considering their complexity, coupling, and frequency. The final predictions are generated using the proposed MTL-CNN-BiLSTM model, optimized with the attention mechanism. The results show that the proposed model significantly reduces error accumulation compared to traditional time-domain analysis methods, achieving a 37.40% reduction in average forecasting error and a 30.73% increase in forecasting efficiency.