Fundamental concepts: Visualization of model performance under various kinds of uncertainty; Further consideration of what is desired from data mining results.
Exemplary techniques: Profit curves; Cumulative response curves; Lift curves; ROC curves.
The previous chapter introduced basic issues of model evaluation and explored the question of what makes for a good model. We developed detailed calculations based on the expected value framework. That chapter was much more mathematical than previous ones, and if this is your first introduction to that material you may have felt overwhelmed by the equations. Though they form the basis for what comes next, by themselves they may not be very intuitive. In this chapter we will take a different view to increase our understanding of what they are revealing.
The expected profit calculation of Equation 7-2 takes a specific set of conditions and generates a single number, representing the expected profit in that scenario. Stakeholders outside of the data science team may have little patience for details, and will often want a higher-level, more intuitive view of model performance. Even data scientists who are comfortable with equations and dry calculations often find such single estimates to be impoverished and uninformative, because they rely on very stringent assumptions (e.g., of precise knowledge of the costs and benefits, or that the models’ estimates of probabilities are accurate). In short, it is ...