2 Iterative estimation of parameters and outof sample prediction.

This is interesting to monitor the evolution of covid-19.

If model is reliable, then also parameter evolution is reliable.

Further we check its out-of-sample prediction of cases since beginning of march. This is more reliable than the above goodness of fit.

2.1 Evolution of parameters (using data since the beginning)

mar 16mar 23mar 30apr 060.000.250.500.751.00
omegaalphabetaDaysPosterior mean and 95% C.I.param

Reaching a peak means α2<0 and β2<0, while disappear of the covid means all parameters less than 0.