In our example, we took the minimum of all the pairwise distances between the data points as the representative of the distance between 2 clusters. This measure of the distance is called a single linkage. Apart from using the minimum, you can use other methods to compute the distance between the clusters.
Let’s see once again the different types of linkages.
- Single Linkage: Here, the distance between 2 clusters is defined as the shortest distance between points in the two clusters
- Complete Linkage: Here, the distance between 2 clusters is defined as the maximum distance between any 2 points in the clusters
- Average Linkage: Here, the distance between 2 clusters is defined as the average distance between every point of one cluster to every other point of the other cluster.
You have to decide what type of linkage should be used by looking at the data. One convenient way to decide is to look at how the dendrogram looks. Usually, a single linkage-type will produce dendrograms which are not structured properly , whereas complete or average linkage will produce clusters which have a proper tree-like structure. You will see later what this means when you run the hierarchical clustering algorithm in Python.
Additional reading
You can read more about the type of linkages here, here and here.
Use the excel file given below to answer the questions that follow:
report an error