Module 10: Clustering

Our first dive into unsupervised learning: finding structure in data that has no labels. This module builds clustering from the ground up — how to group similar samples, how to choose the number of clusters, and how to reason about cluster quality.

What You’ll Learn¶

• k-means and k-means++ centroid initialization

• The elbow method and silhouette analysis for choosing k

• Hard vs. soft (fuzzy) clustering

• Hierarchical clustering and dendrograms

• Density-based clustering with DBSCAN for arbitrary-shaped clusters

Lecture Materials¶

• 📝 Lecture Notes (PDF)

• 💻 Lecture Code (PDF)

• 📓 Chapter notebook: Working with Unlabeled Data — Clustering Analysis

Prerequisites¶

• Modules 1–5 (the scikit-learn workflow, feature scaling, dimensionality reduction)

Reference¶

Raschka & Mirjalili, Machine Learning with PyTorch and Scikit-Learn, Chapter 10.