Chapter 4 Logistic regression

As we saw in Chapters 2 and 3, linear regression assumes that the response variable $Y$ is continuous. In this chapter we will see how logistic regression can deal with a discrete response $Y$. The simplest case is with $Y$ being a binary response, that is, a variable encoding two categories. In general, we assume that we have $X_1,\ldots,X_k$ predictors for explaining $Y$ (multiple logistic regression) and cover the peculiarities for $k=1$ as particular cases.