Time series forecasting of the COVID-19 pandemic: a critical assessment in retrospect DOI Creative Commons
Murat Güngör

Alphanumeric Journal, Journal Year: 2023, Volume and Issue: 11(1), P. 85 - 100

Published: July 12, 2023

The COVID-19 pandemic is perceived by many to have run its course, and forecasting progress no longer a topic of much interest policymakers researchers as it once was. Nevertheless, in order take lessons from this extraordinary two half years, still makes sense critical look at the vast body literature formed thereon, perform comprehensive analyses retrospect. present study directed towards that goal. It distinguished others encompassing all following features simultaneously: (i) time series 10 most affected countries are considered; (ii) for types periods, namely days weeks, analyzed; (iii) wide range exponential smoothing, autoregressive integrated moving average, neural network autoregression models compared means automatic selection procedures; (iv) basic methods benchmarking purposes well mathematical transformations data adjustment taken into account; (v) several test training sizes examined. Our experiments show performance common highly sensitive parameter selection, bound deteriorate dramatically horizon extends, sometimes fails be better than even simplest alternatives. We contend reliableness COVID-19, few weeks ahead, open debate. Policymakers must exercise extreme caution before they make their decisions utilizing forecast such pandemics.

Language: Английский

Carbon emission prediction models: A review DOI
Yukai Jin, Ayyoob Sharifi, Zhisheng Li

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 927, P. 172319 - 172319

Published: April 9, 2024

Language: Английский

Citations

39
Forecasting of daily new lumpy skin disease cases in Thailand at different stages of the epidemic using fuzzy logic time series, NNAR, and ARIMA methods DOI
Veerasak Punyapornwithaya, Orapun Arjkumpa,

Noppawan Buamithup

et al.

Preventive Veterinary Medicine, Journal Year: 2023, Volume and Issue: 217, P. 105964 - 105964

Published: June 16, 2023

Language: Английский

Citations

16
Renewable Energy Forecasting in Turkey: Analytical Approaches DOI Open Access
Mehmet Berke Colak, Erkan Özhan

Journal of Intelligent Systems Theory and Applications, Journal Year: 2025, Volume and Issue: 8(1), P. 25 - 34

Published: March 7, 2025

The growing population and industrialization have resulted in an increased demand for energy, which has worsened environmental problems such as pollution climate change. Renewable energy sources are considered a promising solution due to their benefits limited potential. This study examines the use of neural networks time series analysis predict electricity generation rates from renewable Turkey. We LSTM, NNAR, ELM models, all utilize backpropagation algorithm network forecasting. Additionally, we apply ARIMA, Holt’s trend, linear regression, mean, exponential smoothing models analysis. evaluate performance using mean absolute error root square on training test data. showed that LSTM outperformed ARIMA (1,2,1), (2,2,1), (3,2,1), NNAR methods forecasting accuracy. Although model initially had lowest error, its predictions made it less suitable practical applications. highlights effectiveness predicting sources. (3,2,1) modeling useful optimizing planning management Turkey's future, contributing more sustainable landscape.

Language: Английский

Citations

0
Collateral effects of COVID-19 countermeasures on hepatitis E incidence pattern: a case study of china based on time series models DOI Creative Commons

Yajun Qin,

Haiyang Peng, Jinhao Li

et al.

BMC Infectious Diseases, Journal Year: 2024, Volume and Issue: 24(1)

Published: March 27, 2024

Abstract Background There are abundant studies on COVID-19 but few its impact hepatitis E. We aimed to assess the effect of countermeasures pattern E incidence and explore application time series models in analyzing this pattern. Methods Our pivotal idea was fit a pre-COVID-19 model with data from before outbreak use deviation between forecast values actual reflect countermeasures. analyzed China 2013 2018. evaluated fitting forecasting capability 3 methods outbreak. Furthermore, we employed these construct compare post-COVID-19 forecasts reality. Results Before outbreak, Chinese overall stationary seasonal, peak March, trough October, higher levels winter spring than summer autumn, annually. Nevertheless, were extremely different reality sectional periods congruous others. Conclusions Since pandemic, has altered substantially, greatly decreased. The temporary. anticipated gradually revert

Language: Английский

Citations

3
Comparison of Artificial Intelligence and Machine LearningMethods used in Electric Power System Operation DOI Open Access
Marcel Hallmann,

Robert Pietracho,

Przemlyslaw Komarnicki

et al.

Published: April 5, 2024

The methods of Artificial Intelligence (AI) have been used in the planning and operation power systems for more than 40 years. In recent years, due to development microprocessor data storage technologies, effectiveness this use has greatly increased. This paper provides a systematic overview application AI including Machine Learning (ML) system. potential areas are divided into four blocks classification matrix clustering tasks. Furthermore, acquisition setting parameters ML algorithms presented discussed way considering supervised unsupervised learning methods. Based on this, three complex examples: wind generation forecasting, smart grid security as-sessment (using two methods), automatic system fault detection detail. Summary outlook conclude paper.

Language: Английский

Citations

3
Modeling and Forecasting Dead-on-Arrival in Broilers Using Time Series Methods: A Case Study from Thailand DOI Creative Commons
Chalita Jainonthee, Panneepa Sivapirunthep,

Pranee Pirompud

et al.

Animals, Journal Year: 2025, Volume and Issue: 15(8), P. 1179 - 1179

Published: April 20, 2025

Antibiotic-free (ABF) broiler production plays an important role in promoting sustainable and welfare-oriented poultry farming. However, this system presents challenges, particularly increased susceptibility to stress mortality during transport. This study aimed (i) analyze time series data on the monthly percentage of dead-on-arrival (%DOA) (ii) compare performance various models. Data %DOA from 127,578 transport truckloads recorded between 2018 2024 were aggregated into values. The then decomposed identify trends seasonal patterns. models evaluated included SARIMA, NNAR, TBATS, ETS, XGBoost. These trained using January December 2023, their forecasting accuracy was test 2024. Model assessed multiple error metrics, including MAE, MAPE, MASE, RMSE. results revealed a distinct pattern %DOA. Among models, TBATS ETS demonstrated highest when applied data, with MAPE values 21.2% 22.1%, respectively. considerably lower than those NNAR at 54.4% XGBoost 29.3%. Forecasts for 2025 showed that produced similar can serve as valuable decision-support tool ABF production. By facilitating proactive planning, these help reduce transport-related mortality, improve animal welfare, enhance overall operational efficiency.

Language: Английский

Citations

0
Pakistan CO2 Emission Modelling and Forecasting: A Linear and Nonlinear Time Series Approach DOI Open Access
Kassim Tawiah, Muhammad Daniyal, Moiz Qureshi

et al.

Journal of Environmental and Public Health, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 15

Published: Jan. 31, 2023

Pakistan is considered among the top five countries with highest CO2 emissions globally. This calls for pragmatic policy implementation by all stakeholders to bring finality this alarming situation since it contributes greatly global warming, thereby leading climate change. study an attempt make a comparative analysis of linear time series models nonlinear emission data in Pakistan. These and were used model forecast future values short period. To assess select best these models, we root mean square error (RMSE) absolute (MAE) as performance indicators. The outputs showed that machine learning are other having lowest RMSE MAE values. Based on forecasted value neural network autoregressive model, Pakistan's will be 1.048 metric tons per capita 2028. increasing trend frightening clear warning, suggesting innovative policies must initiated reduce trend. We encourage government price companies entities ton, adapt electricity production from hydro, wind, different sources no CO2, initiate rigorous planting more trees populated areas forest covers, provide incentives companies, organisations, institutions, households come out clean technologies or use those lower ones, fund studies develop less emissions.

Language: Английский

Citations

9
Comparison of Artificial Intelligence and Machine Learning Methods Used in Electric Power System Operation DOI Creative Commons
Marcel Hallmann,

Robert Pietracho,

Przemysław Komarnicki

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(11), P. 2790 - 2790

Published: June 6, 2024

The methods of artificial intelligence (AI) have been used in the planning and operation electric power systems for more than 40 years. In recent years, due to development microprocessor data storage technologies, effectiveness this use has greatly increased. This paper provides a systematic overview application AI, including machine learning (ML) system. potential areas are divided into four blocks classification matrix clustering AI tasks. Furthermore, acquisition setting parameters ML algorithms presented discussed way, considering supervised unsupervised methods. Based on this, three complex examples, being wind generation forecasting, smart grid security assessment (using two methods), automatic system fault detection detail. A summary outlook conclude paper.

Language: Английский

Citations

3
Epidemiological Forecasting Models Using ARIMA, SARIMA, and Holt–Winter Multiplicative Approach for Pakistan DOI Open Access
Muhammad Bilal Riaz, Maqbool Hussain Sial, Saira Sharif

et al.

Journal of Environmental and Public Health, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 8

Published: May 29, 2023

Background of the Study. Statistical models have been extensively used in modeling and forecasting different fields agriculture, economics, social sciences, medical sciences. The transmission some diseases is a serious life threat around globe; therefore, proper assessment need time. Malaria one major life-threatening Pakistan, death cases due to this disease reported during last decade. Methodology. data collected from Ministry Health, Rahim Yar Khan, January 2011 March 2022. Data were analyzed by applying time series for prediction purposes. Diagnostic measures such as RMSE, MAE, MAPE choose best model. Results Discussion. This study aims forecast malaria choosing After comparison, it was concluded that Holt–Winter multiplicative model outperformed ARIMA SARIMA models, with lowest MAPE, MAE compared other models. district Khan forecasted model, month April 2022 2023. From results, minimum number found be 586.75 June maximum 1281.93 October among next ten months. Based on paramount GOP (Govt. Pakistan) enhance vaccination policy erase impacts flatten curve.

Language: Английский

Citations

6
A Comparative Analysis of Traditional SARIMA and Machine Learning Models for CPI Data Modelling in Pakistan DOI Creative Commons
Moiz Qureshi, Arsalan Khan, Muhammad Daniyal

et al.

Applied Computational Intelligence and Soft Computing, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 10

Published: Nov. 7, 2023

Background. In economic theory, a steady consumer price index (CPI) and its associated low inflation rate (IR) are very much preferred to volatile one. CPI is considered major variable in measuring the IR of country. These indices those changes have significance monetary policy decisions. this study, different conventional machine learning methodologies been applied model forecast Pakistan. Methods. Pakistan’s yearly data from 1960 2021 were modelled using seasonal autoregressive moving average (SARIMA), neural network (NNAR), multilayer perceptron (MLP) models. Several forms models compared by employing root mean square error (RMSE), (MSE), absolute percentage (MAPE) as key performance indicators (KPIs). Results. The 20-hidden-layered MLP appeared best-performing for forecasting based on KPIs. Forecasted values 2022 2031 showed an astronomical increase value which unpleasant consumers management. Conclusion. increasing trend observed if not addressed will trigger rising purchasing power, thereby causing higher commodity prices. It recommended that government put vibrant policies place address alarming situation.

Language: Английский

Citations

3