Chapter 7. Model Evaluation

In the previous chapter, we've seen how to create supervised learning methods. We divided our datasets into three subsets—training, validation, and testing. We also used the training dataset to train our models, and in this chapter, we'll use the validation dataset to measure the model performance and to compare different models.

In this chapter, we'll explore different methods for measuring the predictive power of a model.

As we've seen before, there are two kinds of predictive models: regression and classification. In a regression model, the output variable is a numeric variable; in a classification model, the output variable is a categorical variable. We'll start this chapter with cross-validation. After this, we'll measure the performance in regression methods, and then, we'll move on to classification performance.