Title: Reign-and-conquer clustering algorithm Authors:  Miguel de Carvalho - CIDMA, Universidade de Aveiro (Portugal)
Gabriel Martos - Fundacion Universidad Torcuato Di Tella (Argentina) [presenting]
Andrej Svetlosak - University of Edinburgh (United Kingdom)
Abstract: A clustering method is developed that takes advantage of the sturdiness of model-based clustering while attempting to mitigate some of its pitfalls. First, it is noted that standard model-based clustering likely leads to the same number of clusters per margin, which seems a somewhat artificial assumption for a variety of datasets. This issue is tackled by specifying a finite mixture model per margin that allows each margin to have a different number of clusters. Then the multivariate data is clustered using a strategy game-inspired algorithm called Reign-and-Conquer. Second, since the proposed clustering approach only specifies a model for the margins (but leaves the joint unspecified), it has the advantage of being partially parallelizable; hence, the proposed method is computationally appealing as well as more tractable for moderate to high dimensions than a full (joint) model-based clustering approach. A battery of numerical experiments on artificial data indicates an overall good performance of the proposed methods in various scenarios. Real datasets are used to showcase their application in practice.