Python for Finance Cookbook

How it works...

In Step 2, we defined the cross-validation scheme. As there is no inherent order in the data, we used shuffling and specified the random state for reproducibility. Stratification ensured that each fold received a similar ratio of classes in the target variable.

In Step 3, we evaluated the Pipeline created in the Implementing scikit-learn's pipelines recipe using cross_val_score. We passed the estimator (the entire Pipeline), the training data, and the cross-validation scheme as arguments to the function. We could have also provided a number to the cv argument (the default is 5)—if it had been a classification problem, it would have automatically applied the stratified k-fold cross-validation. However, by providing a custom ...