Time series models based on generalized linear models: some further results

Biometrics. 1994 Jun;50(2):506-11.

Abstract

This paper considers the problem of extending the classical moving average models to time series with conditional distributions given by generalized linear models. These models have the advantage of easy construction and estimation. Statistical modelling techniques are also proposed. Some simulation results and an illustrative example are reported to illustrate the methodology. The models will have potential applications in longitudinal data analysis.

MeSH terms

  • Biometry
  • Humans
  • Longitudinal Studies*
  • Models, Statistical*
  • Poisson Distribution
  • Poliomyelitis / epidemiology
  • United States / epidemiology