Chapter 5. Putting Data in its Place – Classification Methods and Analysis
In the previous chapter, we explored methods for analyzing data whose outcome is a continuous variable, such as the purchase volume for a customer account or the expected number of days until cancellation of a subscription service. However, many of the outcomes for data in business analyses are discrete—they may only take a limited number of values. For example, a movie review can be 1–5 stars (but only integers), a customer can cancel or renew a subscription, or an online advertisement can be clicked or ignored.
The methods used to model and predict outcomes for such data are similar to the regression models we covered in the previous chapter. Moreover, sometimes we might want to convert a regression problem into a classification problem: for instance, rather than predicting customer spending patterns in a month, we might be more interested in whether it is above a certain threshold that is meaningful from a business...
