We will use Taiwan's credit card payment defaulters data as an example. This is the same dataset we used earlier, in Chapter 3, Resampling Methods, to build a logistic regression model. In this recipe, we'll build multiple models using different algorithms, and finally, build a stacked ensemble model.

This dataset contains information about credit card clients in Taiwan. This includes information to do with payment defaulters, customers' demographic factors, their credit data, and their payment history. The dataset is provided in GitHub. It is also available from its main source, the UCI ML Repository: https://bit.ly/2EZX6IC.

In our example, we'll use the following supervised algorithms from H2O to build our models:

• Generalized linear model
• Distributed random forest 
• Gradient-boosting machine 
• Stacked ensemble

We'll see how to use these algorithms in Python and learn how to set some of the hyperparameters for each of the algorithms.