A Comparison Between PCA and Hierarchical Clustering

Introduction

Graphical representations of high-dimensional data sets are at the backbone of straightforward exploratory analysis and hypothesis generation. Within the life sciences, two of the most commonly used methods for this purpose are heatmaps combined with hierarchical clustering and principal component analysis (PCA). 

We will use the terminology ‘data set’ to describe the measured data. The data set consists of a number of samples for which a set of variables has been measured. All variables are measured for all samples.