In this chapter, we have looked at the different techniques commonly adopted to evaluate the predictive performances of different algorithms. We looked at how to transform raw data into features, following feature engineering best practices, thereby allowing algorithms to use data that does not have a numeric form, such as categorical variables. We then focused on the techniques needed to correctly evaluate the various components (such as bias and variance) that constitute the generalization error associated with the algorithms, and finally, we learned how to perform the cross validation of the algorithms to improve the training process.
In the next chapter, we will learn how to assess your AI arsenal.
