Bayesian farthest-nearest neighbor for class-imbalanced classification

Jian Xu

doi:10.1016/j.asoc.2026.115681

Bayesian farthest-nearest neighbor for class-imbalanced classification

2026-06-11

0

PRE

AI

DOI：10.1016/j.asoc.2026.115681

Original

Original request for help

Save

Abstract

En 中文

• We propose an updated version of farthest-nearest neighbor query strategy. • We develop a K-free optimization method to determine the optimal number of neighbors. • We design a locally weighted naive Bayes method to estimate class probabilities. • We propose a novel algorithm called Bayesian Farthest-Nearest Neighbor (BFNN). • The experiments on 15 multi-class imbalanced datasets validate its effectiveness.

Bayesian farthest-nearest neighbor for class-imbalanced classification

Abstract

Journal

Organization