Generalised linear models II: Poisson regression
In the previous chapter, we used logistic regression to model binary variables and proportions. In this chapter, we will use another type of generalized linear model, Poisson regression, to model counts. Many medical and epidemiological investigations are concerned with the incidence of diseases in populations. Here, a number of people are followed up and the onset of a disease recorded for each person. The data can then be summarized by counting the total number of occurrences of the disease in different subgroups of subjects sharing the same values of the important explanatory variables. The dependent variable is therefore a count.
KeywordsPoisson Regression Dispersion Parameter Incidence Rate Ratio Deviance Residual Expected Count
Unable to display preview. Download preview PDF.