Affiliation:
1. College of Technological Innovation, Abu Dhabi Campus, Zayed University, FF2-0-056, Abu Dhabi P.O. Box 144534, United Arab Emirates
2. Department of Computer Sciences, Sub Campus Burewala-Vehari, University of Agriculture Faisalabad, Faisalabad 38000, Pakistan
Abstract
The rapidly growing number of COVID-19 infected and death cases has had a catastrophic worldwide impact. As a case study, the total number of death cases in Algeria is over two thousand people (increased with time), which drives us to search its possible trend for early warning and control. In this paper, the proposed model for making a time-series forecast for daily and total infected cases, death cases, and recovered cases for the countrywide Algeria COVID-19 dataset is a two-layer dropout gated recurrent unit (TDGRU). Four performance parameters were used to assess the model’s performance: mean absolute error (MAE), root mean squared error (RMSE), R2, and mean absolute percentage error (MAPE). The results generated with TDGRU are compared with actual numbers as well as predictions with conventional time-series techniques, such as autoregressive integrated moving average (ARIMA), machine learning model of linear regression (LR), and the time series-based deep learning method of long short-term memory (LSTM). The experiment results on different time horizons show that the TDGRU model outperforms the other forecasting methods that deliver correct predictions with lower prediction errors. Furthermore, since this TDGRU is based on a relatively simpler architecture than the LSTM, in comparison to LSTM-based models, it features a significantly reduced number of parameters, a shorter training period, a lower memory storage need, and a more straightforward hardware implementation.
Subject
General Mathematics,Engineering (miscellaneous),Computer Science (miscellaneous)
Reference55 articles.
1. Papastefanopoulos, V., Linardatos, P., and Kotsiantis, S. (2020). COVID-19: A comparison of time series methods to forecast percentage of active cases per population. Appl. Sci., 10.
2. Killeen, B.D., Wu, J.Y., Shah, K., Zapaishchykova, A., Nikutta, P., Tamhane, A., Chakraborty, S., Wei, J., Gao, T., and Thies, M. (2020). A county-level dataset for informing the United States’ response to COVID-19. arXiv.
3. Analyzing the epidemiological outbreak of COVID-19: A visual exploratory data analysis approach;Dey;J. Med. Virol.,2020
4. Prediction and analysis of COVID-19 positive cases using deep learning models: A descriptive case study of India;Arora;Chaos Solitons Fractals,2020
5. Evaluation of E-learning Experience in the Light of the Covid-19 in Higher Education;Abugabah;Procedia Comput. Sci.,2022