K-Means From Scratch | Dataford Interview Questions

Write a function to implement a basic K-means clustering algorithm from scratch.

Given multidimensional points, a cluster count, an iteration limit, and deterministic initial centroids, repeatedly assign each point to its nearest centroid and recompute cluster means. Use squared Euclidean distance and choose the lowest-index centroid when distances tie. If a cluster becomes empty, retain its previous centroid. Return the final centroids and each point's final cluster label.

Input and Output

points, initial_centroids, and centroids are lists of equal-length numeric lists. Return {"centroids": [...], "labels": [...]}. Labels must correspond to the returned centroids.

Constraints

1 <= len(points) <= 1000
1 <= k <= len(points)
1 <= max_iterations <= 100
All points and initial centroids have the same positive dimension
All coordinate values are finite numbers
Initial centroids contains exactly k centroids

Function Signature

def k_means(points, k, max_iterations, initial_centroids):

Your question is K-Means From Scratch. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem