Abstract
The uncertain time series (UTS) is a sequence of uncertain observations
in chronological order. The uncertain autoregressive (UAR) model is one
of the basic UTS models believes the uncertain time series value relies
mainly on it’s historical values linearly. This paper proposes uncertain
interrupted time series (UITS) models aiming at analysing time series
datas with large-scale interventions on the base of uncertain
autoregressive model. The UITS model can reflect the effect of an
intervention and makes prediction about the future in the presence of
intervention. Three types of uncertain interrupted time series models
are introduced in this paper. In addition, residual analysis and
prediction intervals are also proposed. Finally, some numerical examples
are given.