Introduction to hierarchical clustering

Unsupervised Learning in R

Hank Roark

Senior Data Scientist at Boeing

Hierarchical clustering

Number of clusters is not known ahead of time
Two kinds: bottom-up and top-down, this course bottom-up

Simple example

five points

Five clusters

each point is a cluster

Four clusters

four clusters, one cluster with two points

Three clusters

three clusters, two clusters on two points and one of one point

Two clusters

two clusters, one cluster of 3 points and one cluster of 2 points

One cluster

one cluster grouping all points

Hierarchical clustering in R

# Calculates similarity as Euclidean distance 
# between observations
dist_matrix <- dist(x)

# Returns hierarchical clustering model
hclust(d = dist_matrix)

Call:
hclust(d = s)

Cluster method   : complete 
Distance         : euclidean 
Number of objects: 50

Let's practice!

Unsupervised Learning in R