NRCP develops time series regression model for Short term Forecasting of COVID19 Cases

The National Research Council of the Philippines (NRCP) continues to support researches dealing with the social dimensions of the pandemic. Since April 2020, there were already 10 KTOP-COVID short-term projects presented in its KTOP Webinar Series.

Next in line for KTOP webinar is the project which ended on September 6, 2021 that focused on developing a new model in forecasting COVID-19 cases. Titled, “Developing a Time Series Regression Model for Short-term Forecasting of COVID19 Cases” by Dr. Kathleen B. Aviso of De La Salle University, NRCP member of Division of Engineering and Industrial Research, featured an analyses of these COVID scenarios to see their relevance, timeliness, and the necessity for actions. This research is intended to aid in policy making and in helping individuals and organizations adapt new measures in different settings. The study is expected to produce a paper for a KTOP-COVID webinar presentation, a policy recommendation, and a manuscript for publication in the NRCP Research Journal.

The research team of Dr. Aviso saw that there was a need for a new model to be developed to capture the effect of the emergence of new variants and the expected roll-out of vaccination programs. In addition, imperfections in testing coverage may suggest that the true extent of the pandemic is underestimated. Other data routinely gathered by the Department of Health (DOH) may yield clues in forecasting short term trends in pandemic statistics. The capability to predict trends 10-20 days in advance can provide useful guidance on appropriate quarantine levels and required health care resources.

Dr. Aviso’s study developed a hybrid time series regression model that is extended from pure Auto-regressive integrated moving average (ARIMA) models to account for cross-dependencies among variables.  This can be used in forecasting short-term COVID19 infections. The results of this study will be disseminated as the 11th KTOP-COVID webinar on September 30, 2021, 1 month after project completion.