Interpretable Locally Adaptive Nearest Neighbors

TL;DR

This paper introduces a method for learning locally adaptive, interpretable metrics for k nearest neighbors algorithms, enhancing performance and interpretability on synthetic and real-world data.

Contribution

It extends global metric learning to local metrics for kNN, providing a novel, interpretable approach that improves accuracy.

Findings

Improved kNN performance with local metrics

Method is interpretable and adaptable

Effective on both synthetic and real data

Abstract

When training automated systems, it has been shown to be beneficial to adapt the representation of data by learning a problem-specific metric. This metric is global. We extend this idea and, for the widely used family of k nearest neighbors algorithms, develop a method that allows learning locally adaptive metrics. These local metrics not only improve performance but are naturally interpretable. To demonstrate important aspects of how our approach works, we conduct a number of experiments on synthetic data sets, and we show its usefulness on real-world benchmark data sets.