Neural Multivariate Grey Model and Its Applications DOI Creative Commons
Qianyang Li, Xingjun Zhang

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(3), P. 1219 - 1219

Published: Jan. 31, 2024

For time series forecasting, multivariate grey models are excellent at handling incomplete or vague information. The GM(1, N) model represents this group of and has been widely used in various fields. However, constructing a meaningful is challenging due to its more complex structure compared the construction univariate 1). Typically, fitting prediction errors not ideal practical applications, which limits application model. This study presents neural ordinary differential equation (NMGM), new that aims enhance precision models. NMGM employs novel whitening with equations, showcasing higher predictive accuracy broader applicability than previous It can effectively learn features from data samples. In experimental validation, our first predict China’s per capita energy consumption, it performed best both test validation sets, mean absolute percentage (MAPEs) 0.2537% 0.7381%, respectively. optimal results for 0.5298% 1.106%. Then, predicts total renewable lower 0.9566% 0.7896% leading outcomes competing 1.0188% 1.1493%. demonstrate exhibits performance other

Language: Английский

A whale optimization algorithm-based multivariate exponential smoothing grey-holt model for electricity price forecasting DOI
Flavian Emmanuel Sapnken, Ali Khalili Tazehkandgheshlagh, Benjamin Salomon Diboma

et al.

Expert Systems with Applications, Journal Year: 2024, Volume and Issue: 255, P. 124663 - 124663

Published: July 3, 2024

Language: Английский

Citations

8
Improving electricity demand forecasting accuracy: a novel grey-genetic programming approach using GMC(1,N) and residual sign estimation DOI
Flavian Emmanuel Sapnken, Benjamin Salomon Diboma, Ali Khalili Tazehkandgheshlagh

et al.

Grey Systems Theory and Application, Journal Year: 2024, Volume and Issue: 14(4), P. 708 - 732

Published: May 29, 2024

Purpose This paper addresses the challenges associated with forecasting electricity consumption using limited data without making prior assumptions on normality. The study aims to enhance predictive performance of grey models by proposing a novel multivariate convolution model incorporating residual modification and genetic programming sign estimation. Design/methodology/approach research begins constructing demonstrates utilization prediction accuracy exploiting signs forecast residuals. Various statistical criteria are employed assess proposed model. validation process involves applying real datasets spanning from 2001 2019 for annual in Cameroon. Findings hybrid outperforms both non-grey consumption. model's is evaluated MAE, MSD, RMSE, R 2 , yielding values 0.014, 101.01, 10.05, 99% respectively. Results cases real-world scenarios demonstrate feasibility effectiveness combination offers significant improvement over competing models. Notably, dynamic adaptability enhances mimicking expert systems' knowledge decision-making, allowing identification subtle changes demand patterns. Originality/value introduces that incorporates application leveraging residuals represents unique approach. showcases superiority existing models, emphasizing its expert-like ability learn refine rules dynamically. potential extension other fields also highlighted, indicating versatility applicability beyond

Language: Английский

Citations

7
A Nonlinear Multivariate Grey Bernoulli Model for Predicting Innovation Performance in High-Tech Industries DOI
Sandang Guo, Jing Jia, Han Xu

et al.

Communications in Nonlinear Science and Numerical Simulation, Journal Year: 2025, Volume and Issue: unknown, P. 108636 - 108636

Published: Jan. 1, 2025

Language: Английский

Citations

0
A novel time-lag discrete grey Euler model and its application in renewable energy generation prediction DOI
Yong Wang, Rui Yang, Lang Sun

et al.

Renewable Energy, Journal Year: 2025, Volume and Issue: unknown, P. 122785 - 122785

Published: Feb. 1, 2025

Language: Английский

Citations

0
AEPSO: An adaptive learning particle swarm optimization for solving the hyperparameters of dynamic periodic regulation grey model DOI
Gang Hu, Sa Wang,

Bin Shu

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127578 - 127578

Published: April 1, 2025

Language: Английский

Citations

0
Damping accumulative NDAGM(1,N, α) power model and its applications DOI
Ye Li, Chengyun Wang,

Junjuan Liu

et al.

Grey Systems Theory and Application, Journal Year: 2024, Volume and Issue: 14(4), P. 621 - 640

Published: May 8, 2024

Purpose In this essay, a new NDAGM(1,N,α) power model is recommended to resolve the hassle of distinction between old and information, complicated nonlinear traits sequences in real behavior systems. Design/methodology/approach Firstly, correlation aspect sequence screened via grey integrated degree, damped cumulative generating operator index are introduced define model. Then non-structural parameters optimized through genetic algorithm. Finally, pattern utilized for prediction China’s natural gas consumption, contrast with other models. Findings By altering unknown model, theoretical deduction has been carried out on newly constructed It discovered that can be interchanged traditional indicating proposed article possesses strong compatibility. case study, demonstrates superior performance compared benchmark models, which indirectly reflects model’s heightened sensitivity disparities as well its ability handle complex linear issues. Practical implications This paper provides scientifically valid forecast predicting consumption. The results offer foundation formulation national strategies related policies regarding import export. Originality/value primary contribution proposition multivariate accommodates both historical information applicable scenarios. addition, predictive enhanced by employing algorithm search optimal exponent.

Language: Английский

Citations

0
Neural Multivariate Grey Model and Its Applications DOI Creative Commons
Qianyang Li, Xingjun Zhang

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(3), P. 1219 - 1219

Published: Jan. 31, 2024

For time series forecasting, multivariate grey models are excellent at handling incomplete or vague information. The GM(1, N) model represents this group of and has been widely used in various fields. However, constructing a meaningful is challenging due to its more complex structure compared the construction univariate 1). Typically, fitting prediction errors not ideal practical applications, which limits application model. This study presents neural ordinary differential equation (NMGM), new that aims enhance precision models. NMGM employs novel whitening with equations, showcasing higher predictive accuracy broader applicability than previous It can effectively learn features from data samples. In experimental validation, our first predict China’s per capita energy consumption, it performed best both test validation sets, mean absolute percentage (MAPEs) 0.2537% 0.7381%, respectively. optimal results for 0.5298% 1.106%. Then, predicts total renewable lower 0.9566% 0.7896% leading outcomes competing 1.0188% 1.1493%. demonstrate exhibits performance other

Language: Английский

Citations

0