Undirected network-based unsupervised clustering

Author : Hongmei Zhang

Abstract :Development of certain diseases is driven by the concerted work of genetic or epigenetic factors. In unsupervised clustering, it is thus desired to cluster objects based on joint features of such factors. In this piece of work, under the Bayesian framework, we propose an approach to cluster subjects from the lenses of undirected networks, in particular, Gaussian graphs. Pseudo nodes are introduced to assist the selection of tuning parameters in graph construction. Furthermore, since the clustering focuses on the underlying network among the factors (nodes), within cluster variations among subjects with respect to the means of the nodes are allowed and modeled by the Dirichlet process. Simulations and real data applications are implemented to demonstrate the proposed method

Keywords :Bayesian clustering of genetic and epigenetic networks using Gaussian graphs

Conference Name :International Conference on Statistics and Data Science (ICSDS-25)

Conference Place London, UK

Conference Date 15th May 2025