We don't have much time left, and we still need to try at least four models to get the best one and produce the list of customers potentially involved in our profit drop. We are now going to try classification models.

We have already talked about these models in brief, but let me tell you something more structured about them before actually applying them to our data.

Classification models are the ones employed to predict a categorical output. We have already seen that regression models are a good tool when dealing with quantitative output, that is numerical response variables. A typical example is the revenues we were thinking about before. But, what if you don't have a numerical response variable, but categorical variables?