Trouble is, this is the related to the results we want to predict how can we enter an Excel formula for results we don’t have yet? This is where we create our regression equation. Briefly put, the logit is a function that takes a probability of an event as input and returns the logarithm of the odds of that event as output. To create an equation that yields such output will require several calculations. Close to zero means start looking at other schools. Close to one means it probably will get in. The regression equation we hope to create cannot be linear since the permissible output values must fall in the range from zero to one. Our input data represents events that have already occurred, and so 0 means “didn’t get in” and 1 means “got in”. Here, however, we are going to predict a probability value. In a linear regression, an economist might be predicting wages or prices. But there is one very important distinction. So far, it looks a lot like a good old-fashioned linear regression. Our goal is to create an equation with coefficients b 0 through b 4 and then plug in values for gre, gpa, and rank to predict admission. The first few rows of our data look like this: does not have the third possibility, putting you on the waiting list.) (For nit-pickers out there, Hypothetical U. In our hypothetical data, this will be admitted to a hypothetical university program. There are only two possibilities to decide between. Our problem will be the simplest of all categorization problems, a binary categorization. We are going spend more time to obtain our logistic regression by brute force, and the reward for this extra effort will be some greater understanding. If all you want are logistic regression results, there are tools, including the Excel Analysis ToolPack, that will take you there directly. It is important to appreciate that our goal here is to learn about logistic regression. We’re going to gain some insight into how logistic regression works by building a model in Microsoft Excel. In contrast with multiple linear regression, however, the mathematics is a bit more complicated to grasp the first time one encounters it. Logistic regression is one of the most important techniques in the toolbox of the statistician and the data miner.