Fast Estimation of Time-Varying Transmission Rates for Infectious Diseases

Abstract

<p>Modelling and analysis of recurrent infectious disease epidemics often depends on the reconstruction of a time-varying transmission rate from historical reports of cases or deaths. Statistically rigorous estimation methods for time-varying transmission rates exist but are too computationally demanding to apply to a time series longer than a few decades. We present a computationally ecient estimation method that is suitable for very long data sets. Our method, which uses a discrete-time approximation to the SIR model for infectious diseases, is easy to implement and outperforms the classic Fine and Clarkson estimation method.</p>