What is Euclidean distance in terms of machine learning?

It is just a distance measure between a pair of samples p and q in an n-dimensional feature space:

For example, picture it as a "straight, connecting" line in a 2D feature space:

The Euclidean is often the "default" distance used in e.g., K-nearest neighbors (classification) or K-means (clustering) to find the "k closest points" of a particular sample point. Another prominent example is hierarchical clustering, agglomerative clustering (complete and single linkage) where you want to find the distance between clusters.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!