Dendrogram
A dendrogram is a tree-like diagram that records the sequence of merges (or splits) performed by a hierarchical clustering algorithm. It's one of the most intuitive ways to visualize how clusters form step by step.
How to Read a Dendrogram
- Each leaf at the bottom represents a single data point.
- Each merge point (called a "node") shows two clusters being combined.
- The height of a merge represents the distance (or dissimilarity) between the two clusters being joined.
Plotting a Dendrogram in Python
from scipy.cluster.hierarchy import dendrogram, linkage
import matplotlib.pyplot as plt
Z = linkage(data, method="ward")
dendrogram(Z)
plt.title("Dendrogram")
plt.xlabel("Data Points")
plt.ylabel("Distance")
plt.show()Choosing the Number of Clusters
One of the most useful things a dendrogram tells you is how many clusters to pick. By drawing a horizontal line across the tallest vertical lines that aren't crossed by any merge, you can count how many clusters that cut produces.
A good rule of thumb is to cut the dendrogram where the vertical distance between two consecutive merges is largest - this usually indicates the most natural grouping in the data.
Coming Up Next
With clustering covered, the course now shifts to another major unsupervised learning theme - reducing the number of features in your data through Dimensionality Reduction.