Summary

In this chapter, we started exploring unsupervised learning techniques. We focused on cluster analysis to both provide data reduction and data understanding of the observations.

Four methods were introduced: the traditional hierarchical and k-means clustering algorithms, along with PAM, incorporating two different inputs (Gower and Random Forest). We applied these four methods to find a structure in Italian wines coming from three different cultivars and examined the results.

In the next chapter, we will continue exploring unsupervised learning, but instead of finding structure among the observations, we will focus on finding structure among the variables in order to create new features that can be used in a supervised learning problem.