What is the difference between L1 and L2 distance metrics and how does this affect k-NN performance?

Question

Data Science with Python — Hard

What is the difference between L1 and L2 distance metrics and how does this affect k-NN performance?

Accepted Answer

L1 (Manhattan) and L2 (Euclidean) distance metrics differ in how they calculate distances between points. L1 is more robust to outliers and feature differences, making it better for k-NN in high dimensions. L2 penalizes large differences more heavily, affecting k-NN performance.