Predictive Modelling of COVID-19 New Cases in Algeria using An Extreme Learning Machines (ELM)

Djeddou, Messaoud; Hameed, Ibrahim A.; Nejatian, Abolfazl; Loukam, Imed

doi:10.1101/2020.09.28.20203299

Cited by 3 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Djeddou et al [17] predicted new COVID-19 cases in Algeria. The utilized dataset was obtained from the public health database of the Algeria health ministry.…”

Section: Studies Examining African Casesmentioning

confidence: 99%

A Survey of Recent Studies on COVID-19 Outbreak Prediction Using Statistical and Machine Learning Methods

ÇETİN¹,

Abut²

2022

Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi

View full text Add to dashboard Cite

show abstract

“…Djeddou et al [17] predicted new COVID-19 cases in Algeria. The utilized dataset was obtained from the public health database of the Algeria health ministry.…”

Section: Studies Examining African Casesmentioning

confidence: 99%

A Survey of Recent Studies on COVID-19 Outbreak Prediction Using Statistical and Machine Learning Methods

ÇETİN¹,

Abut²

2022

Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

“…An extreme learning machine (ELM) was used to predict the COVID-19 new cases in Algeria. The results showed that the proposed ELM model achieved an adequate level of prediction accuracy with smallest errors [14]. In this study, four artificial neural network models namely (radial basis function neural networks "RBFNN", generalized regression neural networks (GRNN), Extreme learning machine (ELM), and multi-layer perceptron neural network (MLPNN) are proposed for predictive modeling of COVIS-19 new confirmed cases (CNCC) in Algeria.…”

Section: Introductionmentioning

confidence: 99%

Predictive modelling of COVID-19 New Confirmed Cases in Algeria using Artificial Neural Network

Djeddou

Hellal

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

This study investigates the potential of a simple artificial neural network for the prediction of COVID-19 New Confirmed Cases in Algeria (CNCC). Four different ANN models were built (GRNN, RBFNN, ELM, and MLP). The performance of the predictive models is evaluated based on four numerical parameters, namely root mean squared error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (NSE), and Pearson correlation coefficient (R). Taylor diagram was also used to examine the similarities and differences between the observed and predicted values obtained from the proposed models. The results showed the potential of the multi-layer perceptron neural network (MLPNN) which exhibited a high level of accuracy in comparison to the other models.

show abstract

Глобальная мониторинговая информационная система для эпидемиологов (на материале Big Data по COVID-19)

Бычков¹

2021

Клиническая Инфектология И Паразитология

View full text Add to dashboard Cite

Цель. Рассмотреть эффективность паттернов проектирования, разработанных на региональном материале по COVID-19, в плане возможностей использования в глобальной мониторинговой информационной системе для эпидемиологов (здесь – ИС). В качестве паттернов ИС испытать 2-факторную математическую модель, прогнозирующую кумулятивное количество заболевших для регионов Швеция, Финляндия, и 3-факторную математическую модель, прогнозирующую кумулятивное количество заболевших и кумулятивное количество умерших для регионов США, Канада.Материалы и методы. Для 2-факторного, 3-факторного паттернов оценены коэффициент детерминации R2 по матрицам для моделирования и тестирования (валидация при экстраполяции), сила влияния факторов и факторных взаимодействий, рассматриваемых в рамках этих паттернов.Результаты и обсуждение. При разработке и тестировании 2-факторного и 3-факторного паттернов достигнут уровень информативности, стандартный для моих проектов – R2≥0,995. Для 2-факторного паттерна оценка R2 по матрице для моделирования выше, чем таковая по матрице для тестирования (стандартная ситуация). Для 3-факторного паттерна оценка R2 по матрице для тестирования выше, чем таковая по матрице для моделирования (парадоксальная ситуация). Ориентировочно: для 2-факторного паттерна сила влияния факторного взаимодействия составила 20%, для 3-факторного паттерна суммарная сила влияния факторных взаимодействий – 50%. Полученные результаты являются обоснованием учета в ИС максимально возможного количества источников изменчивости (предполагается применение моих ноу-хау, позволяющих минимизировать корреляции между предикторами в матрицах для моделирования при сохранении существенных свойств рассматриваемых систем). Для разработки ИС необходима суперкомпьютерная техника.Выводы. Благодаря синтезу в формате, родственном метаанализу, ИС функционирует на основе использования генеральной многофакторной нелинейной эмпирической математической модели, описывающей при R2≥0,995 Big Data по всем странам. При получении прогнозов по каждому региону учитываются данные по всем остальным регионам. Назначение ИС – обоснованные оперативные прогнозирование динамики рассматриваемых пандемий, оценка, мониторинг эффективности мер, принимаемых в регионах, оптимизация этих мер с учетом ряда условий. Эффективность ИС пропорциональна количеству рассматриваемых источников изменчивости. Purpose.To examine the effectiveness oftypalpatternsdevelopedontheregional COVID-19 material in terms of the possibilities of using in Global Monitoring Information System for Epidemiologists (here – “IS”). As IS patterns, to test 2-factorial mathematical model that prognosticates cumulative the number of illnesses (total cases) for Sweden / Finland and 3-factorial mathematical model that prognosticates cumulative the number of illnesses (total cases) and cumulative death rate (total deaths) for USA / Canada.Materials and methods. For 2-factorial and 3-factorial patterns, the determination coefficient R2 on the matrix for modelling and the matrix for testing (validation at extrapolation) and the force of the influence of the factors and factorial interactions considered within these patterns were estimated. Results and discussion. In the development and testing of 2-factorial and 3-factorial patterns, the level of informativeness that is standard for my projects has been reached – R2≥0.995. For 2-factorial pattern, the estimation of R2 on the modelling matrix is higher than those on the testing matrix (usual situation). For 3-factorial pattern, the estimation of R2 on the testing matrix is higher than those on the modelling matrix (paradoxical situation). Approximately: for 2-factorial pattern the force of influence of factor interaction was 20%, for 3-factorial pattern total force of influence of all factor interactions was 50%. The results justify taking into account the maximum possible number of the sources of variation in IS (it is assumed that my know-how will be used to minimize the correlation between predictors in the modelling matrix while maintaining the essential properties of the system under consideration). The development of IS requires supercomputer technics.Conclusions. Thanks to the synthesis in the format similar to meta-analysis, IS operates on the basis of using general multifactorial nonlinear empirical mathematical model that describes at R2≥0.995 Big Data across all countries. At prognosticating for each region take into account data for all the rest regions. The designation of IS – reasonable operative prognosticating of the dynamics of the pandemics under consideration, assessment, monitoring of the effectiveness of the measures taken in the regions, optimization of these measures taking into account a number of conditions. The effectiveness of IS is proportional to the number of the effects (the sources of variation) considered.

show abstract

Predictive Modelling of COVID-19 New Cases in Algeria using An Extreme Learning Machines (ELM)

Cited by 3 publications

References 20 publications

A Survey of Recent Studies on COVID-19 Outbreak Prediction Using Statistical and Machine Learning Methods

A Survey of Recent Studies on COVID-19 Outbreak Prediction Using Statistical and Machine Learning Methods

Predictive modelling of COVID-19 New Confirmed Cases in Algeria using Artificial Neural Network

Глобальная мониторинговая информационная система для эпидемиологов (на материале Big Data по COVID-19)

Contact Info

Product

Resources

About