Box-Jenkins modelling of some viral infectious diseases

Abstract

During the last few years Box-Jenkins models have become of increasing importance in such fields as economics and industry. They have been used for forecasting and for detecting relations between different time series. Similar applications are also relevant in epidemiology. Since many people who are concerned with the analysis of medical data are not familiar with this subject, a short non-technical introduction to Box-Jenkins models is given. The model-building process is demonstrated in some detail using monthly case reports of the two seasonal endemic diseases chicken-pox and mumps, and the relation between these two time series is investigated. It is found that both diseases may be represented by simple models which have basically the same statistical structure and that they are related at time-lag 0 but not at other time-lags.