Approximation of Estimates in the Susceptible-Infectious-Removed Epidemic Model

Abstract

In this note, a simple approximation for the maximum likelihood estimates of infection and removal parameters used in the Susceptible-Infectious-Removed (SIR) epidemic model is presented. This approximation can be applied when the numbers of susceptible and infected individuals are observable only at discrete points in time. Since, in such cases, a closed form of the likelihood function is generally too complicated to obtain, the proposed approximation method represents an important advance. Simulation results show that the method yields approximations quite close to the maximum likelihood estimates obtained under continuous observation.