Question

When forecasting the number of infected people from COVID-19, what kind of forecasting model you think would be appropriate? For example, exponential smoothing, linear trend, or moving average. Please choose one or more and elaborate the reasons that you think your selected forecasting model is appropriate.

Finally, what is your forecasted number of infected people from COVID-19 based on your selected forecasting model? Please explain your reasons / procedures to get your forecasted number.

Answer 1

Answer:

While forecasting the number of infected people from COVID-19, the kind of forecasting model that would be appropriate is exponential smoothing model.

The data of the infected people from COVID-19 shows an exponential increase. Exponential forecasting is used for forecasting the infected people as it shows good forecast accuracy in a short period of time as compared to the other forecasting techniques. Moreover, exponential smoothing model takes into consideration the various trends and seasonal forecasting patterns including the additive pattern and the multiplicative pattern and the combination of both.

The total confirmed COVID 19 cases till 30/04/2020 is approximately 3.2 million around the world.

The forecast has been carried out by taking into consideration the 10 days of interval for which the forecast trend was established starting from 31/01/2020. The forecast started with the multiplicative trend exponential smoothing model since the data available for the forecast was limited during the starting of the forecast. For every ten days of forecast we established a mean value for the number of confirmed cases and compared the forecast result at the end of the tenth day with that of the mean value. The deviation was noticed for every ten days of forecast. For the next ten days of forecast, the previous data of the tenth day was taken as a reference and the new mean was established for the next ten days. The forecast for the new tenth day is confirmed with the new mean and the relationship with the actual data and prediction is calculated along with the determination of the variation between the new mean and the new tenth day data. It has been observed that as we moved over the ten days of forecast, for every new interval of 10 days, the forecast has shown the precise data for the confirmed cases of COVID 19 from the previous interval of 10 days. The deviation of the data and the mean has decreased considerably over the successive data obtained for the period of 10 days.