Multiple Closed-Form Local Metric Learning for K-Nearest Neighbor Classifier

TL;DR

This paper introduces a computationally efficient framework for learning multiple Mahalanobis distance metrics in closed-form, enhancing kNN classifier performance without iterative procedures.

Contribution

It presents a novel closed-form method for local metric learning, overcoming the limitations of iterative and computationally expensive approaches.

Findings

Achieves improved kNN classification accuracy

Reduces computational complexity of metric learning

Enables learning multiple metrics simultaneously

Abstract

Many researches have been devoted to learn a Mahalanobis distance metric, which can effectively improve the performance of kNN classification. Most approaches are iterative and computational expensive and linear rigidity still critically limits metric learning algorithm to perform better. We proposed a computational economical framework to learn multiple metrics in closed-form.