Determining an effective short term COVID-19 prediction model in ASEAN countries

Area of Science:

• Epidemiology
• Data Science
• Time Series Analysis

Background:

• Accurate short-term forecasting of disease spread, like COVID-19, is crucial for public health preparedness.
• Model selection and data trend characteristics pose significant challenges in demand prediction.
• Existing forecasting models may not fully capture the complexities of cumulative case data.

Purpose of the Study:

• To identify the most efficient model for short-term COVID-19 cumulative case forecasting in ASEAN countries.
• To evaluate the performance of Holt's method, Wright's modified Holt's method, and the unreplicated linear functional relationship model (ULFR).
• To determine the optimal smoothing parameters for accurate prediction.

Main Methods:

• Collected cumulative COVID-19 case data from the Worldometers database for ASEAN countries.
• Applied Holt's method, Wright's modified Holt's method, and the unreplicated linear functional relationship model (ULFR).
• Utilized Nash-Sutcliffe efficiency (NSE) and R-squared for model validation and selection, comparing them to traditional R-squared.

Main Results:

• Holt's method and Wright's modified Holt's method yielded identical results due to the absence of missing data.
• The unreplicated linear functional relationship model (ULFR) demonstrated superior prediction ability compared to Holt's method.
• One-day ahead forecasting proved most efficient, with performance validated by NSE across 1, 3, and 7-day forecasts.

Conclusions:

• The unreplicated linear functional relationship model (ULFR) is an efficient forecasting model for identifying trends in cumulative COVID-19 cases.
• Model performance is influenced by data volume and the chosen smoothing parameters.
• ULFR offers a more accurate alternative to traditional methods like Holt's for short-term epidemiological forecasting.