Abstract
The dynamics of coronavirus cases is proposed to be modeled using switching regression whose switching points are unknown. The stepwise process of constructing the regression in time is described. The dynamics of the number of coronavirus cases in Ukraine is analyzed.