The experimental results show that weekly scale look-back windows (LBW) provide an optimal balance between accuracy and computational cost for day-ahead horizons and the DL model NHiTS, when integrating exogenous features and a 7-day LBW, improves forecasting accuracy by 50.8%, outperforming all other deep learning, machine learning, and statistical forecasting models.
Abstract
Accurate day-ahead cooling load forecasting is a time series forecasting problem that is essential for optimizing the scheduling and energy efficiency of chilled water plants (CWPs). While previous studies have explored different forecasting models and input features, the combined effects of key factors such as historical context length (including regressor lags), sampling resolution, and exogenous inputs remain insufficiently examined for real industrial data. This study examines, under varying regressors and data granularity, the amount of historical context that provides stable and reliable forecasts with low computational cost. The experimental results show that (i) weekly scale look-back windows (LBW) provide an optimal balance between accuracy and computational cost for day-ahead horizons, (ii) hourly inputs often match or exceed finer-resolution accuracy and reveal distinct compute scaling: deep-learning (DL) models remain comparatively efficient as resolution increases, whereas the machine learning (ML) model, XGBoost, grows steeply, (iii) incorporating weather and calendar features enhances prediction accuracy around weekends and holidays. The DL model NHiTS, when integrating exogenous features and a 7-day LBW, improves forecasting accuracy by 50.8%, outperforming all other deep learning, machine learning, and statistical forecasting models by achieving a MASE of 1.88, compared to 3.82 for the baseline. These findings provide practical insights for smart-building and industrial applications, guiding practitioners such as building-energy engineers and facility managers in selecting optimal historical data lengths and parameter combinations to improve forecasting accuracy.
Accurate forecasting of agricultural commodity prices is a critical task for market participants and policymakers, particularly in contexts characterized by multiple sources of uncertainty. Meteorological conditions are often considered potential drivers of price dynamics and are therefore frequently included in predictive models. This study examines the case of fine durum wheat by comparing Machine Learning and Deep Learning approaches under different feature configurations, with and without meteorological variables. Using long-term monthly data, we evaluate eXtreme Gradient Boosting (XGBoost) and Long Short-Term Memory (LSTM) networks within a unified out-of-sample forecasting framework. The empirical results show that, at the monthly frequency, including weather-related variables does not yield systematic improvements in forecasting accuracy. In several cases, models that exclude meteorological inputs achieve comparable or superior performance while reducing complexity. In particular, XGBoost trained on price-based covariates yields the most robust and accurate predictions, whereas LSTM models capture long-term trends but do not consistently benefit from climatic information. These findings highlight the importance of careful feature selection and suggest that the predictive value of weather variables is highly dependent on data granularity and market structure.
Predicting electrical demand in distribution systems is a fundamental problem for the efficient operation of smart grids, especially under scenarios of high temporal variability. This study compares two multi-step forecasting strategies for 24-h horizons: a direct 24→24 strategy and a recursive strategy based on sequential 24→1 predictions. Four machine learning and deep learning architectures are evaluated: LSTM, N-HiTS, U-Net, and LightGBM, using real data from electrical feeders belonging to distribution systems in the equatorial region of Ecuador. The methodology includes constructing time windows, non-overlapping train/validation/test partitioning for evaluation, consistent normalization, and comparative analysis using MAE, RMSE, and MAPE metrics. The results show that the direct 24→24 strategy achieves the best overall performance, with LSTM standing out with an approximate MAPE of 4.12%. However, the recursive strategies exhibit greater stability in the face of atypical patterns observed during holidays and weekends. Furthermore, U-Net demonstrates competitive performance in both accuracy and temporal robustness, while LightGBM stands out for its computational efficiency. It is concluded that the selection of a forecasting strategy depends on the required balance between overall accuracy, temporal stability, and computational cost.
Erik Fernando Mendez-Garces, David Buldain, M. Comech· Energies· 0 citations
Accurate meteorological forecasting is essential for agricultural planning, irrigation management, and environmental decision support. This study conducts a comparative evaluation of recurrent and hybrid deep learning architectures for multivariate forecasting of reference evapotranspiration ($ET_0$), vapour pressure deficit (VPD), wind speed, and the sine and cosine components of wind direction. The analysis utilizes 134,376 hourly observations from Ioannina, Greece, spanning January 2011 to April 2026, sourced from ERA5 via the OpenMeteo Historical Weather API. Single and multi-layer GRU and LSTM networks are compared with hybrid 1D-CNN-GRU and 1D-CNN-LSTM models for two forecasting tasks: a 24-hour next-day forecast and a 168-hour week-ahead forecast. Performance is evaluated using normalized root mean squared error, the coefficient of determination, and a composite Weighted Quotient Score (WQS). The most effective purely recurrent models are a 64-unit LSTM for the 24-hour horizon, with a WQS of 0.816755, and a 1024-unit GRU for the 168-hour horizon, with a WQS of 0.779465. The hybrid CNN-GRU models achieved the highest overall scores of 0.827535 and 0.782863 for the 24-hour and 168-hour horizons, but with additionally more number of units respectively to LSTM models, while the CNN-LSTM models yield nearly identical results with substantially fewer parameters. Compared to the corresponding recurrent baselines, the hybrid models improve WQS by 1.22--1.63\% at 24 hours and by 0.44--0.45\% at 168 hours, indicating that convolutional feature extraction is more beneficial for short-term forecasting.
The intermittent and volatile characteristics of new energy generation, together with the increasing demand for stable power supply in intelligent industrial systems, make accurate forecasting a critical issue for grid dispatch and electromagnetic energy management. This study systematically reviews the technological evolution of time series analysis methods for wind and photovoltaic (PV) power forecasting and establishes a comparative framework covering classical statistical models, intelligent learning algorithms, and hybrid modeling strategies. Based on two years of operational data collected from an actual wind farm and PV station in East China, the forecasting performance of ARIMA, exponential smoothing, Support Vector Regression (SVR), Long Short-Term Memory (LSTM) networks, and Transformer architectures is comprehensively evaluated, while hybrid approaches based on Empirical Mode Decomposition (EMD) are further investigated. The results demonstrate that model selection should jointly consider forecasting horizon, data characteristics, and computational constraints. Classical statistical methods remain robust under stable operating conditions but are less effective in capturing extreme fluctuations, whereas deep learning approaches exhibit superior capability in modeling long-range temporal dependencies despite reduced interpretability. Decomposition-based hybrid strategies achieve a more balanced performance across diverse scenarios and show enhanced robustness under extreme weather conditions. The study further proposes a structured model selection guideline by matching data characteristics with operational requirements, providing theoretical support for forecasting system design in renewable-energy-driven power networks and offering useful references for electromagnetic energy utilization and intelligent industrial applications.
M. Song, C. Yang, Z. Heng et al.· Advanced Electromagnetics· 0 citations
—This paper investigates electricity load forecasting using machine learning models enhanced with advanced optimization techniques. Six regression-based models—Gradient Boosting, LightGBM, ExtraTrees, Random Forest, Decision Tree, and Long Short-Term Memory (LSTM)—are evaluated on two real-world datasets from Panama City and Tetouan City, across hourly and 10-minute temporal resolutions. Results demonstrate that tree-based ensemble models, particularly the ExtraTreesRegressor, consistently outperform LSTM-based deep learning approaches. A key contribution is the development of an Enhanced Harris Hawks Optimization (EHHO) algorithm, incorporating adaptive parameter control and type-specific parameter handling. EHHO significantly improves hyperparameter tuning efficiency, enabling the ExtraTreesRegressor to achieve state-of-the-art forecasting accuracy. The EHHO-optimized ExtraTreesRegressor attains a Mean Absolute Percentage Error (MAPE) of 0.30% for Tetouan City and 1.47% for Panama City using 10-minute resolution data. The analysis reveals that higher temporal granularity contributes up to 65% improvement in forecasting performance compared to hourly data. These findings challenge the prevailing view of deep learning dominance in time-series forecasting and establish new accuracy benchmarks for electricity load prediction. The proposed methodology holds strong potential for practical deployment in grid operation, demand response, and renewable energy integration, supporting the development of more efficient and resilient energy systems
Ahmed Rashed Almesmari, M. Al-Betar, S. Makhadmeh· Journal of Advances in Infor...· 0 citations
For smart grid power system planning and operation, short-term load forecasting is crucial. Important decisions including determining system safety, scheduling fuel, economically dispatching electricity, and selling energy can be aided by accurate day-ahead estimates. However, due to its reliance on external variables like weather, the process is intricate and computationally intensive. In order to address this issue, the paper proposes an LSTGR-based architecture that systematically enhances STLF in Smart Grids. The input characteristics are first scaled correctly using data normalisation. A hybrid feature selection method combining XGB and RF is utilised to determine the most essential features. Afterwards, RFE is employed to eliminate superfluous attributes. To facilitate learning, a hybrid deep learning model is trained using the updated dataset. This model combines LSTM and GRU. Using assessment criteria such as MAPE, MAE, MSE, and RMSE, the results demonstrate that the LSTGR model outperforms other models. With an RMSE of only 1.8%, the model clearly excels at producing accurate predictions. All things considered, the model successfully improves the reliability of forecasts while being computationally efficient. Because of this, it is an excellent option for smart grid applications in the actual world.
L. Jayavani, Banoth Ashwini, Kolkur Swabhavika et al.· 2026 7th International Confe...· 0 citations
