Clustering methodologies

There are many ways to conduct segmentation. There are various methodologies that one can use to influence how small, big, or distinct the cluster constituents are from each other. Without delving much into the methodology, let's first look at what we mean by the distance between clusters and how it impacts on the results. In Figure 7.10, we have four imaginary lines, L1-L4, which we will use to understand the distance between clusters. The x-axis shows the customer IDs and the y-axis shows the distance between cluster centroids. The higher the distance, the further the clusters are from each other. At L1, we can see that we don't have a relationship between dissimilar customers. Customers 5 and 6 and 2, 3 have matching values and hence they are paired together, but they don't form a cluster with any other (duplicate/non-matching), customers...