Unsupervised learning

Unsupervised Learning

Finding groups in data without labels.

4 lessons~35 min2 formulas5 questions

Not started yet

Quiz mastery0%

What you’ll take away

Clustering finds groups without labels; judge it with silhouette or known labels.
k-means alternates assign/update and suits round clusters of similar size.
Hierarchical clustering builds a dendrogram you can cut at any level.
DBSCAN finds arbitrarily shaped dense clusters and flags noise.

Major concepts (syllabus)

Clustering, k-means, hierarchical clustering, density-based clustering

Lessons

J = k = 1 \sum K x_{i} \in C_{k} \sum ∥ x_{i} - μ_{k} ∥^{2}

Within-cluster sum of squares.

s (i) = \frac{b ( i ) - a ( i )}{max { a ( i ) , b ( i )}}

Near 1 means well clustered.

Clustering: Grouping unlabelled examples so that similar ones end up together.
DBSCAN: Density-based clustering that finds clusters of any shape and marks low-density points as noise.
k-means: Clustering that alternates assigning points to the nearest centroid and recomputing centroids.
Silhouette score: Clustering quality per point: (b − a) / max(a, b), from −1 to 1.