Fig 1.
Cumulative new COVID-19 cases in Australia for the whole span of the data.
Two structural break points in the dynamics of the COVID-19 population are evident: a major break point on 27 March 2020, and a minor break point on 18 April 2020.
Table 1.
MLE of parameters.
Fig 2.
Cumulative new COVID-19 cases in Australia for three categories consisting of: (i) training data spanning from 1−20 March 2020 (in black), test data spanning from 21−26 March 2020 (in red), and (iii) predicted values over the period of test data (in blue) with MAPE = 1.53%.
Fig 3.
Cumulative new COVID-19 cases in Australia for the whole span of the data (in logarithmic scale) compared with the predicted values for the scenario in which the lockdown restrictions had not been implemented.
It shows the exponential growth of the COVID-19 population before the impact of lockdown restrictions on 27 March 2020 (marked in green).
Fig 4.
The log-likelihood function of partial observations in terms of the two parameters λ and p.
This plot illustrates the identifiability of the estimates provided in Table 1.
Fig 5.
MAPE of predictions vs. size of training data for a range of size of training data (test data) varying from 20−25 (6−1).
Table 2.
MLE of parameters.
Fig 6.
Cumulative new COVID-19 cases in Australia for three categories: (i) training data spanning 1 March–3 April 2020 (in black), test data spanning 4−17 April 2020 (in red), and (iii) predicted values over the period of test data (in blue) with MAPE = 0.43%.
The first break point on 27 March 2020 is marked in green.
Table 3.
MLE of parameters.
Fig 7.
Cumulative new COVID-19 cases in Australia for three categories: (i) training data spanning 1 March–24 April 2020 (in black), test data spanning 25 April–15 June 2020 (in red), and (iii) predicted values over the period of test data (in blue) with MAPE = 0.26%.
The two break points are marked in green.