Incremental clustering based on Wasserstein distance between histogram models

Xiaotong Qian, Guénaël Cabanes, Parisa Rastin, Mohamed Alae Guidani, Ghassen Marrakchi, Marianne Clausel, Nistor Grozavu

Abstract: Highlights•Fast, flexible framework for static/dynamic clustering over sliding windows.•Uses histogram models to represent clusters with arbitrary distributions.•High degree of flexibility in selecting the clustering algorithm to apply.•Achieves efficient, high quality results with Wasserstein-based statistical tests.