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

Djeddou, M.; A. Hameed, I.; Nejatian, A.; Loukam, I.

In this research, an extreme learning machine (ELM) is proposed to predict the new COVID-19 cases in Algeria. In the present study, public health database from the Algeria health ministry has been used to train and test the ELM models. The input parameters for the predictive models include Cumulative Confirmed COVID-19 Cases (CCCC), Calculated COVID-19 New Cases (CCNC), and Index Day (ID). The predictive accuracy of the seven models has been assessed via several statistical parameters. The results showed that the proposed ELM model achieved an adequate level of prediction accuracy with smallest errors (MSE= 0.16, RMSE=0.4114, and MAE= 0.2912), and highest performances (NSE = 0.9999, IO = 0.9988, R2 = 0.9999). Hence, the ELM model could be utilized as a reliable and accurate modeling approach for predicting the new COVIS-19 cases in Algeria.