Comparative analysis of Gated Recurrent Units (GRU), long Short-Term memory (LSTM) cells, autoregressive Integrated moving average (ARIMA), seasonal autoregressive Integrated moving average (SARIMA) for forecasting COVID-19 trends

Alexandria Engineering Journal, Journal Year: 2022, Volume and Issue: 61(10), P. 7585 - 7603

Published: Jan. 6, 2022

Several machine learning and deep models were reported in the literature to forecast COVID-19 but there is no comprehensive report on comparison between statistical models. The present work reports a comparative time-series analysis of techniques (Recurrent Neural Networks with GRU LSTM cells) (ARIMA SARIMA) country-wise cumulative confirmed, recovered, deaths. Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM) cells based (RNN), ARIMA SARIMA trained, tested, optimized trends COVID-19. We deployed python optimize parameters which include (p, d, q) representing autoregressive moving average terms model additional seasonal are denoted by (P, D, Q). Similarly, for RNN models' (number layers, hidden size, rate number epochs) deploying PyTorch framework. best was chosen lowest Mean Square Error (MSE) Root Squared (RMSE) values. For most data countries, learning-based outperformed models, an RMSE values that 40 folds less than But some countries (ARIMA, Further, we emphasize importance various factors such as age, preventive measures healthcare facilities etc. play vital role rapid spread pandemic.

Language: Английский

---

An Overview of Forecast Analysis with ARIMA Models during the COVID-19 Pandemic: Methodology and Case Study in Brazil

Mathematics, Journal Year: 2023, Volume and Issue: 11(14), P. 3069 - 3069

Published: July 12, 2023

This comprehensive overview focuses on the issues presented by pandemic due to COVID-19, understanding its spread and wide-ranging effects of government-imposed restrictions. The examines utility autoregressive integrated moving average (ARIMA) models, which are often overlooked in forecasting perceived limitations handling complex dynamic scenarios. Our work applies ARIMA models a case study using data from Recife, capital Pernambuco, Brazil, collected between March September 2020. research provides insights into implications adaptability predictive methods context global pandemic. findings highlight models’ strength generating accurate short-term forecasts, crucial for an immediate response slow down disease’s rapid spread. Accurate timely predictions serve as basis evidence-based public health strategies interventions, greatly assisting management. model selection involves automated process optimizing parameters autocorrelation partial plots, well various precise measures. performance chosen is confirmed when comparing forecasts with real reported after forecast period. successfully both recovered COVID-19 cases across preventive plan phases Recife. However, model’s observed extend future. By end period, error substantially increased, it failed detect stabilization deceleration cases. highlights challenges associated such under-reporting recording delays. Despite these limitations, emphasizes potential while emphasizing need further enhance long-term predictions.

Language: Английский

---

A proficient approach to forecast COVID-19 spread via optimized dynamic machine learning models

Scientific Reports, Journal Year: 2022, Volume and Issue: 12(1)

Published: Feb. 14, 2022

This study aims to develop an assumption-free data-driven model accurately forecast COVID-19 spread. Towards this end, we firstly employed Bayesian optimization tune the Gaussian process regression (GPR) hyperparameters efficient GPR-based for forecasting recovered and confirmed cases in two highly impacted countries, India Brazil. However, machine learning models do not consider time dependency data series. Here, dynamic information has been taken into account alleviate limitation by introducing lagged measurements constructing investigated models. Additionally, assessed contribution of incorporated features prediction using Random Forest algorithm. Results reveal that significant improvement can be obtained proposed In addition, results highlighted superior performance GPR compared other (i.e., Support vector regression, Boosted trees, Bagged Decision tree, Forest, XGBoost) achieving averaged mean absolute percentage error around 0.1%. Finally, provided confidence level predicted based on showed predictions are within 95% interval. presents a promising shallow simple approach predicting

Language: Английский

---

Application of machine learning in the prediction of COVID-19 daily new cases: A scoping review

Heliyon, Journal Year: 2021, Volume and Issue: 7(10), P. e08143 - e08143

Published: Oct. 1, 2021

COVID-19 has produced a global pandemic affecting all over of the world. Prediction rate spread and modeling its course have critical impact on both health system policy makers. Indeed, making depends judgments formed by prediction models to propose new strategies measure efficiency imposed policies. Based nonlinear complex nature this disorder difficulties in estimation virus transmission features using traditional epidemic models, artificial intelligence methods been applied for spread. importance machine deep learning approaches spreading trend, present study, we review studies which used these predict number cases COVID-19. Adaptive neuro-fuzzy inference system, long short-term memory, recurrent neural network multilayer perceptron are among mostly regard. We compared performance several Root means squared error (RMSE), mean absolute (MAE), R

Language: Английский

---

Predicting the impact of the third wave of COVID-19 in India using hybrid statistical machine learning models: A time series forecasting and sentiment analysis approach

Computers in Biology and Medicine, Journal Year: 2022, Volume and Issue: 144, P. 105354 - 105354

Published: Feb. 26, 2022

Language: Английский

---

Computational Machine Learning Approach for Flood Susceptibility Assessment Integrated with Remote Sensing and GIS Techniques from Jeddah, Saudi Arabia

Remote Sensing, Journal Year: 2022, Volume and Issue: 14(21), P. 5515 - 5515

Published: Nov. 2, 2022

Floods, one of the most common natural hazards globally, are challenging to anticipate and estimate accurately. This study aims demonstrate predictive ability four ensemble algorithms for assessing flood risk. Bagging (BE), logistic model tree (LT), kernel support vector machine (k-SVM), k-nearest neighbour (KNN) used in this zoning Jeddah City, Saudi Arabia. The 141 locations have been identified research area based on interpretation aerial photos, historical data, Google Earth, field surveys. For purpose, 14 continuous factors different categorical examine their effect flooding area. dependency analysis (DA) was analyse strength predictors. comprises two input variables combination (C1 C2) features sensitivity selection. under-the-receiver operating characteristic curve (AUC) root mean square error (RMSE) were utilised determine accuracy a good forecast. validation findings showed that BE-C1 performed best terms precision, accuracy, AUC, specificity, as well lowest (RMSE). performance skills overall models proved reliable with range AUC (89–97%). can also be beneficial flash forecasts warning activity developed by disaster

Language: Английский

---

A predictive analytics framework for sensor data using time series and deep learning techniques

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: 36(11), P. 6119 - 6132

Published: Jan. 18, 2024

Abstract IoT devices convert billions of objects into data-generating entities, enabling them to report status and interact with their surroundings. This data comes in various formats, like structured, semi-structured, or unstructured. In addition, it can be collected batches real time. The problem now is how benefit from all this gathered by sensing monitoring changes temperature, light, position. paper, we propose a predictive analytics framework constructed on top open-source technologies such as Apache Spark Kafka. focuses forecasting temperature time series using traditional deep learning methods. analysis prediction tasks were performed Autoregressive Integrated Moving Average (ARIMA), Seasonal (SARIMA), Long Short-Term Memory (LSTM), novel hybrid model based Convolution Neural Network (CNN) LSTM. purpose paper determine whether recently developed learning-based models outperform algorithms the data. empirical studies conducted reported demonstrate that models, specifically LSTM CNN-LSTM, exhibit superior performance compared traditional-based algorithms, ARIMA SARIMA. More specifically, average reduction error rates obtained CNN-LSTM substantial when other indicating superiority learning. Moreover, CNN-LSTM-based exhibits higher degree closeness actual values LSTM-based model.

Language: Английский

---

AI-powered COVID-19 forecasting: a comprehensive comparison of advanced deep learning methods

Osong Public Health and Research Perspectives, Journal Year: 2024, Volume and Issue: 15(2), P. 115 - 136

Published: March 28, 2024

Objectives: The coronavirus disease 2019 (COVID-19) pandemic continues to pose significant challenges the public health sector, including that of United Arab Emirates (UAE). objective this study was assess efficiency and accuracy various deep-learning models in forecasting COVID-19 cases within UAE, thereby aiding nation’s authorities informed decision-making.Methods: This utilized a comprehensive dataset encompassing confirmed cases, demographic statistics, socioeconomic indicators. Several advanced deep learning models, long short-term memory (LSTM), bidirectional LSTM, convolutional neural network (CNN), CNN-LSTM, multilayer perceptron, recurrent (RNN) were trained evaluated. Bayesian optimization also implemented fine-tune these models.Results: evaluation framework revealed each model exhibited different levels predictive precision. Specifically, RNN outperformed other architectures even without optimization. Comprehensive perspective analytics conducted scrutinize dataset.Conclusion: transcends academic boundaries by offering critical insights enable UAE deploy targeted data-driven interventions. model, which identified as most reliable accurate for specific context, can significantly influence decisions. Moreover, broader implications research validate capability techniques handling complex datasets, thus transformative potential healthcare sectors.

Language: Английский

---

Incidence and prediction of cutaneous leishmaniasis cases and its related factors in an endemic area of Southeast Morocco: time series analysis

Acta Tropica, Journal Year: 2025, Volume and Issue: unknown, P. 107579 - 107579

Published: March 1, 2025

Language: Английский

---

An effective dimensionality reduction approach for short-term load forecasting

Electric Power Systems Research, Journal Year: 2022, Volume and Issue: 210, P. 108150 - 108150

Published: June 11, 2022

Language: Английский

---