Estimating the probability of New Zealand regions being free from COVID-19 using a stochastic SEIR model

Rafal Bogacz

is the (which was not peer-reviewed) The copyright holder for this preprint current-cases/. Form the available data we extracted the number of cases in each DHB on each day between 25 March and 18 April 2020. Subsequently we used the fitted model to predict probability of the population being virus free P (E t = 0, I t = 0|∆R 1:t ). To avoid the influence of the prior probability (that was set arbitrarily) on the estimate, for each DHB model was additionally run on the first week of data 10 times, i.e. the model was estimating the probability on the basis of extended sequence of new cases composed of: [∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:7 , ∆R 1:t ].