Unsupervised Learning

October 5, 2026 machine-learning

Handwritten notes on unsupervised learning: clustering algorithms, dimensionality reduction, anomaly detection, and density estimation.

These notes explore learning from unlabeled data — finding hidden structure and patterns:

Clustering: K-Means, hierarchical clustering, DBSCAN, silhouette scores, and elbow method.
Dimensionality Reduction: PCA, t-SNE, eigenvalue decomposition, and variance preservation.
Anomaly Detection: Gaussian-based detection, isolation forests, and threshold selection.
Density Estimation: Gaussian Mixture Models, Expectation-Maximization, and latent variable models.