Modelling and predicting the spread of Coronavirus (COVID-19) infection in NUTS-3 Italian regions
share
This paper represents a frst attempt to model and predict the spatio-temporal distribution of Coronavirus (COVID-19) infections at NUTS-3 regional level in Italy. The statistical endemic-epidemic multivariate areal count time-series model introduced by Paul and Held (2011), and further refined by Meyer and Held (2014), is employed. Despite being very preliminary, the results are promising with respect to the possibility to gain useful insights about the spatial and temporal characteristics of both the endemic and epidemic behaviours of this new infectious disease. Moreover, the result also show that the model has the potential to provide good forecasts of the number of infections at local level while controlling for under-reporting.
READ FULL TEXT