1: \begin{abstract}\noindent

2: Count data with excessive zeros are often encountered when modelling

3: infectious disease occurrence. The degree of zero inflation can vary over

4: time due to non-epidemic periods as well as by age group or region.

5: The existing endemic-epidemic modelling framework (\emph{aka} HHH)

6: lacks a proper treatment for surveillance data with excessive zeros as it

7: is limited to Poisson and negative binomial distributions.

8: In this paper, we propose a multivariate zero-inflated endemic-epidemic

9: model with random effects to extend HHH.

10: % Parameters in the zero-inflation part can be interpreted in terms of odds ratio, 

11: % whereas parameters in the HHH part has the same interpretation 

12: % after a zero-inflation adjustment as in the HHH model. 

13: Parameters of the new zero-inflation and the HHH part of the model can be

14: estimated jointly and efficiently via (penalized) maximum likelihood

15: inference using analytical derivatives.

16: A simulation study confirms proper convergence and

17: coverage probabilities of confidence intervals.

18: Applying the model to measles counts in the 16 German states, 2005--2018,

19: shows that the added zero-inflation improves probabilistic

20: forecasts.

21: 

22: \bigskip\noindent\textit{Keywords}:

23: zero inflation, multivariate time series, epidemic modelling, seasonality, measles

24: \end{abstract}

25: