Time series models for tracking and forecasting COVID
Authors: Andrew Harvey and Paul Kattuman
Abstract: Time series models are developed for predicting future values of a variable that when cumulated is subject to an unknown saturation level. Such models are relevant for many disciplines, but attention is focused on the spread of epidemics with applications for coronavirus. These growth curve based models can be assessed by standard statistical test procedures. Estimates of the reproduction number are obtained as a by-product. Applications are discussed.