Clustering is the first unsupervised method in the course: it finds structure in data with no labels to learn from. In finance it groups assets into peers, sorts trading days into market regimes, and compresses a universe of names into a handful of representatives. k-Means is the algorithm interviewers ask about first, because it is simple, fast, and built on a clean objective you can write down and optimize by hand. This lesson covers the objective, the iterative algorithm, works a full run on concrete numbers, shows how to choose the number of clusters, and contrasts k-means with the alternatives. We tie each idea to trading and to the interview.
Table of Contents
Start Your Interview Preparations!
Already have an account? Log in!