Decision Analytics: Microsoft® Excel®

2. Logistic Regression

In This Chapter

The Rationale for Logistic Regression

The Distribution of the Residuals

Using Logistic Regression

Maximizing the Log Likelihood

The Rationale for Log Likelihood

The Statistical Significance of the Log Likelihood

One of the tools of decision analytics is called logistic regression. This method actually isn’t any sort of regression at all: It uses maximum likelihood estimates (which are optimized probabilities, as you’ll see later in this chapter) instead of correlations to arrive at its solutions. The use of the word regression appears to be due to the similarity of the equations that result. For example, in ordinary, least squares multiple regression you might arrive at an equation that looks like this: ...