Evolutionary weighted mean based framework for generalized median computation with application to strings

Abstract

A new general framework for generalized median approximation is proposed based on the concept of weighted mean of a pair of objects. It can be easily adopted for different application domains like strings, graphs or clusterings, among others. The framework is validated for strings showing its superiority over the state-of-the-art. © 2012 Springer-Verlag Berlin Heidelberg.