Variational autoencoder-based learning intrinsic periodic-trend representations of power load series for short-term forecasting

Energy Reports, Год журнала: 2025, Номер 13, С. 6584 - 6595

Опубликована: Июнь 1, 2025

Язык: Английский

---

Self-Similar Traffic Prediction Algorithm for Satellite Network Based on Dual Decomposition and Neural Network

Computer Networks, Год журнала: 2025, Номер unknown, С. 111432 - 111432

Опубликована: Июнь 1, 2025

Язык: Английский

---

Hybrid modeling approaches for agricultural commodity prices using CEEMDAN and time delay neural networks

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Ноя. 4, 2024

Abstract Improving the forecasting accuracy of agricultural commodity prices is critical for many stakeholders namely, farmers, traders, exporters, governments, and all other partners in price channel, to evade risks enable appropriate policy interventions. However, traditional mono-scale smoothing techniques often fail capture non-stationary non-linear features due their multifarious structure. This study has proposed a CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise)-TDNN (Time Delay Neural Network) model non-linear, series. evaluated its suitability comparison three major EMD (Empirical Decomposition) variants (EMD, Complementary EMD) benchmark (Autoregressive Integrated Moving Average, Non-linear Support Vector Regression, Gradient Boosting Machine, Random Forest TDNN) models using monthly wholesale oilseed crops India. Outcomes from this investigation reflect that CEEMDAN-TDNN hybrid have outperformed on basis evaluation metrics under consideration. For model, an average improvement RMSE (Root Mean Square Error), Relative MAPE (Mean Absolute Percentage Error) values been observed be 20.04%, 19.94% 27.80%, respectively over variant-based counterparts 57.66%, 48.37% 62.37%, stochastic machine learning models. The CEEMD-TDNN demonstrated superior performance predicting directional changes series compared Additionally, forecasts generated by assessed Diebold-Mariano test, Friedman Taylor diagram. results confirm alternative models, providing distinct advantage.

Язык: Английский

---

Improved Bacterial Foraging Optimization Algorithm with Machine Learning-Driven Short-Term Electricity Load Forecasting: A Case Study in Peninsular Malaysia

Algorithms, Год журнала: 2024, Номер 17(11), С. 510 - 510

Опубликована: Ноя. 6, 2024

Accurate electricity demand forecasting is crucial for ensuring the sustainability and reliability of power systems. Least square support vector machines (LSSVM) are well suited to handle complex non-linear load series. However, less optimal regularization parameter Gaussian kernel function in LSSVM model have contributed flawed accuracy random generalization ability. Thus, these parameters need be chosen appropriately using intelligent optimization algorithms. This study proposes a new hybrid based on optimized by improved bacterial foraging algorithm (IBFOA) short-term daily Peninsular Malaysia. The IBFOA sine cosine equation addresses limitations fixed chemotaxis constants original (BFOA), enhancing its exploration exploitation capabilities. Finally, LSSVM-IBFOA constructed mean absolute percentage error (MAPE) as objective function. comparative analysis demonstrates model, achieving highest determination coefficient (R2) 0.9880 significantly reducing average MAPE value 28.36%, 27.72%, 5.47% compared deep neural network (DNN), LSSVM, LSSVM-BFOA, respectively. Additionally, exhibits faster convergence times BFOA, highlighting practicality forecasting.

Язык: Английский

---

A novel method for long-term power demand prediction using enhanced data decomposition and neural network with integrated uncertainty analysis: A Cuba case study

Applied Energy, Год журнала: 2024, Номер 372, С. 123864 - 123864

Опубликована: Июль 9, 2024

Язык: Английский

---