Importance of careful data tracking for short-term prediction of COVID-19 case and death numbers from official resources

During the COVID-19 outbreak, governments, scientists, health workers, and numerous people worked on strategies or solutions for halting disease propagation. Unfortunately, need monitoring is steeply increasing, taking necessary restrictive actions currently unavoidable. Due to lack of epidemiological data constantly changing numbers, constructing less error-prone predictive models reliable mathematical near future will help make better legal prevention strategies. 
 In this study, daily eleven countries between 01/21/2020-05/02/2020 01/21/2020-06/17/2020 were used forecast number events by using different forecasting models. Best fit chosen after analysis present numbers with Auto-Regressive Integrated Moving Average(ARIMA), Brown’s LES, Holt’s LES based MAPE values. The study showed least best-fit short-term predictions analyzing two datasets demonstrated that changed updates among selected countries. Investigation USA (Holt’s MAPE=7,7 ARIMA(2,2,0) MAPE=4,8 case ARIMA (2,0,0) MAPE=5,7 ARIMA(1,2,0) MAPE=3,4 death numbers), Turkey (ARIMA(2,0,0) MAPE=4,0 Brown's MAPE=1,7 ARIMA(2,1,1) MAPE=0,9 ARIMA(0,2,0) Brazil (Holt's MAPE=6,2 ARIMA(1,0,1) MAPE=36,4 MAPE=3,2 MAPE=2,8 Russia (ARIMA(1,2,0) MAPE=6,8 MAPE=3,5 ARIMA(1,1,1) MAPE=3,7 numbers) at same time flow, updating caused alterations in model selection, which results changes predictions.
 indicate more than one statistical has superiority current approaches fluctuations should be taken into account when construct create management already complicated exhausting pandemic. Thus, policies restrictions against spread might successful considering adjusted providing accurate results.