What changes when a KNN classifier uses weights='distance' instead of uniform weights?
Classic MLLevel: IntermediateNeighbor influence
What changes when a KNN classifier uses weights='distance' instead of uniform weights?
Hints
- Neighbor membership may stay the same.
- The vote weights change.
Answer
Nearer neighbors receive more influence, typically inverse-distance weighting, instead of equal votes. This changes the local decision rule and can increase sensitivity to very close noisy points.
Source and licence: scikit-learn 1.7.2 — neighbors · BSD-3-Clause · answer: Quantoros editorial