Summary

In this chapter, we looked at using probabilistic linear models to predict a qualitative response with three methods: logistic regression, discriminant analysis, and MARS. Additionally, we began the process of using ROC charts in order to explore model selection visually and statistically. We also briefly discussed the model selection and trade-offs that you need to consider. In future chapters, we will revisit the breast cancer dataset to see how more complex techniques perform.