Summary

In this chapter, you learned about cross-validation, which is one of the most important resampling methods. It results in the best estimation of model performance on independent data. This chapter covered the basics of cross-validation and its two different variations, leave-one-out and k-fold, along with a comparison of them.

Next, we covered the Keras wrapper with scikit-learn, which is a very helpful tool that allows scikit-learn methods and functions that perform cross-validation to be easily applied to Keras models. Following this, you were shown a step-by-step process of implementing cross-validation in order to evaluate Keras deep learning models using the Keras wrapper with scikit-learn.

Finally, you learned that cross-validation estimations of model performance can be used to decide between different models for a particular problem or to decide which parameters (or hyperparameters) should be used for a particular model. You practiced using cross-validation for...