Back to Search
Start Over
Deep Learning Algorithms for Forecasting COVID-19 Cases in Saudi Arabia.
- Source :
- Applied Sciences (2076-3417); Feb2023, Vol. 13 Issue 3, p1816, 23p
- Publication Year :
- 2023
-
Abstract
- In the recent past, the COVID-19 epidemic has impeded global economic progress and, by extension, all of society. This type of pandemic has spread rapidly, posing a threat to human lives and the economy. Because of the growing scale of COVID-19 cases, employing artificial intelligence for future prediction purposes during this pandemic is crucial. Consequently, the major objective of this research paper is to compare various deep learning forecasting algorithms, including auto-regressive integrated moving average, long short-term memory, and conventional neural network techniques to forecast how COVID-19 would spread in Saudi Arabia in terms of the number of people infected, the number of deaths, and the number of recovered cases. Three different time horizons were used for COVID-19 predictions: short-term forecasting, medium-term forecasting, and long-term forecasting. Data pre-processing and feature extraction steps were performed as an integral part of the analysis work. Six performance measures were applied for comparing the efficacy of the developed models. LSTM and CNN algorithms have shown superior predictive precision with errors of less than 5% measured on available real data sets. The best model to predict the confirmed death cases is LSTM, which has better RMSE and R 2 values. Still, CNN has a similar comparative performance to LSTM. LSTM unexpectedly performed badly when predicting the recovered cases, with RMSE and R 2 values of 641.3 and 0.313, respectively. This work helps decisionmakers and health authorities reasonably evaluate the status of the pandemic in the country and act accordingly. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 20763417
- Volume :
- 13
- Issue :
- 3
- Database :
- Complementary Index
- Journal :
- Applied Sciences (2076-3417)
- Publication Type :
- Academic Journal
- Accession number :
- 161819591
- Full Text :
- https://doi.org/10.3390/app13031816