In later chapters, we'll discuss methods for how to choose optimal values for alpha, gamma, and epsilon in more detail. For now, we'll use the values we have and test different values against each other manually. One of the most straightforward options is a cross-validation method, such as a grid search, which can be done programmatically.
Recall that hyperparameter tuning in machine learning is the process of finding the hyperparameters for a model (such as the depth or number of nodes of a decision tree) that will get the best performance for that model:
![](/api/v2/epubs/9781789345803/files/assets/f704ea35-39ed-47f6-be62-815e893ee5fd.png)
As we mentioned earlier, the hyperparameter values are not determined ...