Summary

In this chapter, we extended most of the models and methods learned earlier in the book. The chapter began with a detailed example of housing data, and we carried out the visualization and pre-processing. The principal component method helps in reducing data, and the variable clustering method also helps with the same task. Linear regression models, neural networks, and the regression tree were then introduced as methods that will serve as base learners. Bagging, boosting, and random forest algorithms are some methods that helped to improve the models. These methods are based on homogeneous ensemble methods. This chapter then closed with the stacking ensemble method for the three heterogeneous base learners.

A different data structure of censored observations will be the topic of the next chapter. Such data is referred to as survival data, and it commonly appears in the study of clinical trials.