Forecasting COVID-19 new cases in South Africa using ARIMA and LSTM time-series analysis methods

dc.contributor.authorZulu, Sindisiwe
dc.date.accessioned2023-07-20T13:34:27Z
dc.date.available2023-07-20T13:34:27Z
dc.date.issued2023
dc.descriptionA research report submitted in partial fulfillment of the requirements for the degree of Master of Science to the Faculty of Science, School of Computer Science and Applied Mathematics, University of the Witwatersrand, Johannesburg, 2023
dc.description.abstractThe COVID-19 disease continues to threaten the lives of human beings both in South Africa and globally. Its impact has been felt in practically every facet of life, resulting in a new standard of living globally. In an effort to reduce the spread of the virus, a model that can accurately estimate new cases of COVID-19 in the future is required. The daily cases in South Africa were modeled in this study utilizing Auto-Regressive Integrated Moving Average (ARIMA) model and Long Short-Term Memory (LSTM) model. Forecasts for the last 15 days of data were generated using the models. Two model performance metrics, namely the MAE and RMSE were used to find the model that can best forecast future COVID-19 cases. The LSTM model was the superior model resulting in the smallest MAE and RMSE values. This study highlights the predictive capability of LSTM model, which can be used as a tool for forecasting the spread of COVID-19 to assist the health officials and government authorities in better preparing and being informed when managing the spread of the disease.
dc.description.librarianNG (2023)
dc.facultyFaculty of Science
dc.identifier.urihttps://hdl.handle.net/10539/35739
dc.language.isoen
dc.schoolSchool of Computer Science and Applied Mathematics
dc.titleForecasting COVID-19 new cases in South Africa using ARIMA and LSTM time-series analysis methods
dc.typeDissertation
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
MSc_Report_Sindisiwe_Zulu_2030273.pdf
Size:
1.07 MB
Format:
Adobe Portable Document Format
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
2.43 KB
Format:
Item-specific license agreed upon to submission
Description:
Collections