Unsupervised Learning
Find structure in data that has no response to predict, and face the consequence squarely: with no y there is no held-out error, so every choice you make has to be defended some other way.
Sign in to take quizzes, earn XP, and unlock stages as you reach 90% mastery.
Principal Components and Variance
25 min · 100 XPThe direction of maximum variance, the constraint that makes the question well posed, the proportion of variance explained, and why scaling is not optional.
Open lesson →Sign in to take the 3-question quiz.
K-Means and Local Optima
25 min · 100 XPThe within-cluster variation, why the exact problem is intractable, the argument that the algorithm must converge, and why a single run cannot be trusted.
Open lesson →Sign in to take the 3-question quiz.
Hierarchical Clustering and Linkage
25 min · 100 XPAgglomerative fusion, reading a dendrogram correctly, what each linkage does differently, and the decisions that change the answer with no test to settle them.
Open lesson →Sign in to take the 3-question quiz.