A novel initialisation based on hospital-resident assignment for the $$k$$-modes algorithm

Abstract This paper presents a new way of selecting an initialisation for the $$k$$ k -modes algorithm that allows notion game theoretic fairness classic initialisations, namely those by Huang and Cao, do not. Our method utilises hospital-resident assignment problem to find set initial cluster centroids which we compare with two classical methods -modes: original presented next most popular Cao co-authors. To highlight merits our proposed method, stages analysis are presented. It is demonstrated often able offer computational speed-up order $$50\%$$ 50 % . Improved clustering, in terms commonly used cost-function, was witnessed several cases can be $$10\%$$ 10 , particularly more complex datasets.