Generalizing fuzzy k-nearest neighbor classifier using an OWA operator with a RIM quantifier (OWARIM-FKNN)

This repository offers the code for $${\color{blue} Generalizing \space fuzzy \space k-nearest \space neighbor \space classifier \space using \space an \space OWA \space operator \space} $$ $${\color{blue} with \space a \space RIM \space quantifier \space}$$

Introduction:
OWARIM-FKNN is an enhanced version of the fuzzy k-nearest neighbor (FKNN) classifier, desinged to improve classification accuracy while dealing with outliers and data uncertainty issues. The method uses an ordered weighted averaging (OWA) operator and RIM quantifier to create multiple local mean vectors for each class to generate class representative psudo-nearest neighbors.

Matlab functions:
The code of the Md-FKNNreg algorithm (owarim_fknn.m) and related required functions owamatrix.m and RIM.m are included. In addition two example cases are also given.

Reference: KUmbure, M.M. and Luukka, P. Generalizing fuzzy k-nearest neighbor classifier using an OWA operator with a RIM quantifier (2025), Expert Systems with Applications, 282, 127795 https://doi.org/10.1016/j.eswa.2025.127795