A mathematical model and forecast for the coronavirus disease COVID-19 in Ukraine (M3)
(A.B. Alyokhin, B.V. Burkynskyi, A.N. Grabovoi, V.A, Dilenko, N.I. Khumarova)
11/10/2020. No F-228.
The first full M3 weekly forecast of the COVID-19 epidemic in Ukraine published first a week ago on October 5, 2020 (see https://www.facebook.com/ab.alyokhin/posts/186389343003969). Today we publish an extended analysis of the accuracy of this forecast.
The results of the analysis of the weekly forecast compiled using this model are showing in table. 1–2 and fig. 1–12.
Table 1a shows the calculated and actual values of the main cumulative indicators of the COVID-19 epidemic in Ukraine for the 7-day forecast period. The absolute and relative errors of daily forecasts, as well as the average absolute error of MAE and the average absolute error in percentage of MAPE, including the same estimate of MAPE * growing the result, the forecast as a whole.
For the first time, we provide similar estimates of the forecasting accuracy of all major daily indicators (see Table 1b).
Table 2 shows the same data for case fatality rates I (TC), I (CC) and the IP progress indicator.
For a long time, in our forecasts for different countries in the world, we achieved high accuracy in forecasting cumulative and other synthetic indicators of the COVID-19 epidemic. In the M3 model, the emphasis is on forecasting daily indicators. As we have repeatedly noted, and as it follows from all the charts containing the actual trajectories of the daily indicators of the COVID-19 epidemic, these indicators are extremely variable and especially difficult to forecast. We set ourselves the task to improve the technique of forecasting such indicators.
The average absolute percentage errors (MAPE * estimates) of the forecast for the main cumulative indicators of the epidemic are at the following level (Table 1a):
- total number of infected — 1.14%;
- total number of deaths — 0.59%;
- total number of recovered — 0.73%;
- number of active cases — 2.52%.
The average absolute percentage errors (MAPE * estimates) of the forecast for the main daily indicators of the epidemic are at the following level (Table 1b):
- total number of infected people — 14.66%;
- total number of deaths — 12.02%;
- total number of recovered — 25.86%;
- number of active cases — 33.96%.
The errors in forecasting the most important derived indicators of the COVID-19 epidemic in Ukraine (MAPE * estimates) are at the following level (see Table 2):
- mortality rate I (TC) — 0.56%;
- mortality rate I (CC) — 1.18%;
- IP progress indicator — 1.68%,
Visually, the accuracy of forecasting the main indicators of the epidemic can assess using the diagrams in Fig. 1–12. On all charts, the forecast trajectories of the forecasted indicators for the 7-day period marked in dark blue, and the trajectories of the actual values of the same indicators are red.
In fig. 1–4 show the actual and calculated trajectories of the main cumulative indicators of the COVID-19 epidemic in Ukraine for the entire observation period, including the 7-day period.
As you can see, over the past week, the reality turned out to be slightly worse than the prognosis in terms of the number of infected (Fig. 1), deaths (Fig. 2) and recovered (Fig. 3).
Diagrams Fig. 5–8 reflect the actual and calculated trajectories of the main daily indicators of the COVID-19 epidemic in Ukraine for the entire observation period, including the forecast period.
Among the daily indicators, only the growth rate of deaths markedly deviated from the forecasted upward, demonstrating an unusually low level of variability, constantly remaining at a high level.
Diagrams Fig. 9–12 reflect the actual and calculated trajectories of synthetic indicators of the epidemic for the entire observation period, including the 7-day forecast period:
- average absolute increase in the number of infected;
- lethality rate I (TC);
- lethality rate I (CC);
- IP progress indicator.
Among the specified synthetic indicators of the epidemic, only the progress indicator reacted negatively to the trends of the past week, deviating from the more optimistic forecast for the worse.
The relatively high accuracy of the forecast once again confirms that the extremely negative trends in the development of the COVID-19 epidemic have become stable.
Sources of statistical data:
Our materials also:
Accuracy of our forecasts:
Publications on case fatality rates and progress indicator