Modeling and forecasting electricity consumption amid the COVID-19 pandemic: Machine learning vs. nonlinear econometric time series models

Sustainable Cities and Society, Journal Year: 2023, Volume and Issue: 98, P. 104860 - 104860

Published: Aug. 15, 2023

Accurately modelling and forecasting electricity consumption remains a challenging task due to the large number of statistical properties that characterize this time series such as seasonality, trend, sudden changes, slow decay autocrrelation function, among many others. This study contributes literature by using comparing four advanced econometrics models, machine learning deep models1 analyze forecast during COVID-19 pre-lockdown, lockdown, releasing-lockdown, post-lockdown phases. Monthly data on Qatar's total has been used from January 2010 December 2021. The empirical findings demonstrate both econometric models are able capture most important features characterizing consumption. In particular, it is found climate change based factors, e.g temperature, rainfall, mean sea-level pressure wind speed, key determinants terms forecasting, results indicate autoregressive fractionally integrated moving average three state markov switching with exogenous variables outperform all other models. Policy implications energy-environmental recommendations proposed discussed.

Language: Английский

---

Innovative framework for accurate and transparent forecasting of energy consumption: A fusion of feature selection and interpretable machine learning

Applied Energy, Journal Year: 2024, Volume and Issue: 366, P. 123314 - 123314

Published: April 29, 2024

The study presents a novel framework integrating feature selection (FS) and machine learning (ML) techniques to forecast inland national energy consumption (EC) in the United Kingdom across all sources. This innovative strategically combines three FS approaches with five interpretable ML models using Shapley Additive Explanations (SHAP), dual goal of enhancing accuracy transparency EC predictions. By meticulously selecting most pertinent features from diverse features—including meteorological conditions, socioeconomic parameters, historical patterns different primary fuels—the proposed enhances robustness forecasting model. is achieved through benchmarking approaches: ensemble filter, wrapper, hybrid filter-wrapper. In addition, we introduce filter FS, synthesizing outcomes multiple base methods make well-informed decisions about retention. Experimental results underscore efficacy both wrapper filter-wrapper models, ensuring process remains comprehensible while utilizing manageable number (four eight). experimental indicate that subsets are usually selected for each combined approach not only demonstrates framework's capability provide accurate forecasts but also establishes it as valuable tool policymakers analysts.

Language: Английский

---

Enhancing Building Energy Efficiency with IoT-Driven Hybrid Deep Learning Models for Accurate Energy Consumption Prediction

Sustainability, Journal Year: 2024, Volume and Issue: 16(5), P. 1925 - 1925

Published: Feb. 26, 2024

Buildings remain pivotal in global energy consumption, necessitating a focused approach toward enhancing their efficiency to alleviate environmental impacts. Precise prediction stands as linchpin optimizing efficiency, offering indispensable foresight into future demands critical for sustainable environments. However, accurately forecasting consumption individual households and commercial buildings presents multifaceted challenges due diverse patterns. Leveraging the emerging landscape of Internet Things (IoT) smart homes, coupled with AI-driven solutions, promising avenues overcoming these challenges. This study introduces pioneering that harnesses hybrid deep learning model prediction, strategically amalgamating convolutional neural networks’ features long short-term memory (LSTM) units. The granularity IoT-enabled meter data, enabling precise forecasts both residential spaces. In comparative analysis against established models, proposed consistently demonstrates superior performance, notably exceling predicting weekly average usage. study’s innovation lies its novel architecture, showcasing an unprecedented capability forecast holds significant promise guiding tailored management strategies, thereby fostering optimized practices buildings. demonstrated superiority underscores potential serve cornerstone driving utilization, invaluable guidance more energy-efficient future.

Language: Английский

---

A Set Pair Analysis Method for Assessing and Forecasting Water Conflict Risk in Transboundary River Basins

Water Resources Management, Journal Year: 2023, Volume and Issue: 38(2), P. 775 - 791

Published: Dec. 19, 2023

Language: Английский

---

Statistical Comparison of Time Series Models for Forecasting Brazilian Monthly Energy Demand Using Economic, Industrial, and Climatic Exogenous Variables

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(13), P. 5846 - 5846

Published: July 4, 2024

Energy demand forecasting is crucial for effective resource management within the energy sector and aligned with objectives of Sustainable Development Goal 7 (SDG7). This study undertakes a comparative analysis different models to predict future trends in Brazil, improve methodologies, achieve sustainable development goals. The evaluation encompasses following models: Seasonal Autoregressive Integrated Moving Average (SARIMA), Exogenous SARIMA (SARIMAX), Facebook Prophet (FB Prophet), Holt–Winters, Trigonometric Seasonality Box–Cox transformation, ARMA errors, Trend, components (TBATS), draws attention their respective strengths limitations. Its findings reveal unique capabilities among models, excelling tracing seasonal patterns, FB demonstrating its potential applicability across various sectors, Holt–Winters adept at managing fluctuations, TBATS offering flexibility albeit requiring significant data inputs. Additionally, investigation explores effect external factors on consumption, by establishing connections through Granger causality test conducting correlation analyses. accuracy these assessed without exogenous variables, categorized as economical, industrial, climatic. Ultimately, this seeks add body knowledge prediction, well allow informed decision-making planning policymaking and, thus, make rapid progress toward SDG7 associated targets. paper concludes that, although achieves best accuracy, most fit model, considering residual autocorrelation, it predicts that Brazil will approximately 70,000 GWh 2033.

Language: Английский

---

Short- and long-term forecasting for building energy consumption considering IPMVP recommendations, WEO and COP27 scenarios

Applied Energy, Journal Year: 2023, Volume and Issue: 339, P. 120980 - 120980

Published: March 22, 2023

Language: Английский

---

A novel Seasonal Fractional Incomplete Gamma Grey Bernoulli Model and its application in forecasting hydroelectric generation

Energy, Journal Year: 2024, Volume and Issue: 290, P. 130257 - 130257

Published: Jan. 6, 2024

Language: Английский

---

A novel damped conformable fractional grey Bernoulli model and its applications in energy prediction with uncertainties

Applied Mathematical Modelling, Journal Year: 2024, Volume and Issue: 130, P. 94 - 118

Published: March 6, 2024

Language: Английский

---

Multi-Step Ahead Forecasting of the Energy Consumed by the Residential and Commercial Sectors in the United States Based on a Hybrid CNN-BiLSTM Model

Sustainability, Journal Year: 2023, Volume and Issue: 15(3), P. 1895 - 1895

Published: Jan. 19, 2023

COVID-19 has continuously influenced energy security and caused an enormous impact on human life social activities due to the stay-at-home orders. After Omicron wave, economy system are gradually recovering, but uncertainty remains virus mutations that could arise. Accurate forecasting of consumed by residential commercial sectors is challenging for efficient emergency management policy-making. Affected geographical location long-term evolution, time series prominent temporal spatial characteristics. A hybrid model (CNN-BiLSTM) based a convolution neural network (CNN) bidirectional long short-term memory (BiLSTM) proposed extract features, where features captured CNN layer, extracted BiLSTM layer. Then, recursive multi-step ahead strategy designed forecasting, grid search employed tune hyperparameters. Four cases 24-step in United States given evaluate performance model, comparison with 4 deep learning models 6 popular machine 12 evaluation metrics. Results show CNN-BiLSTM outperforms all other four cases, MAPEs ranging from 4.0034% 5.4774%, improved 0.1252% 49.1410%, compared models, which also about 5 times lower than 5.9559% average. It evident prediction accuracy great potential sectors.

Language: Английский

---

DESCINet: A hierarchical deep convolutional neural network with skip connection for long time series forecasting

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 228, P. 120246 - 120246

Published: April 28, 2023

Language: Английский

---