7Generalized Linear Models

In Chapter 4 we considered regression models where the response variable c07-i0001 depends linearly on several explanatory variables c07-i0002 In this case, c07-i0003 was a quantitative variable; that is, it could take on any real value, and the regressors – which could be quantitative or qualitative – affected only its mean.

In this chapter we shall consider more general situations in which the regressors affect the distribution function of c07-i0004 However, to retain parsimony, it is assumed that this distribution depends on them only through a linear combination c07-i0005 where the c07-i0006s are unknown.

We shall first consider the situation when c07-i0007 is a 0–1 variable.

7.1 Binary response regression

Let c07-i0008 be a 0–1 ...

Get Robust Statistics, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.