A0476
Title: Informed partition models for dependent random partitions
Authors: Sally Paganin - The Ohio State University (United States) [presenting]
Garritt Page - Brigham Young University (United States)
Fernando Quintana - Pontificia Universidad Catolica de Chile (Chile)
Abstract: Model-based clustering is a powerful tool often used to discover hidden structure in data by grouping observational units that exhibit similar response values. Recently, clustering methods have been developed that allow the inclusion of an initial partition of the data informed by expert opinions, starting from a probability distribution on the space of partitions. Then, using some similarity criteria, partitions different from the initial one are down-weighted, i.e., they are assigned reduced probabilities. A different perspective is taken, and the probability that each unit follows the initial partition via auxiliary variables is modeled. The informed partition model provides flexibility to include varying levels of uncertainty to any subset of the partition (i.e., locally weighted prior information). Additionally, it can accommodate settings with multiple dependent partitions, such as temporal or multi-view data. Theoretical properties of the proposed construction are explored, which can be useful for prior elicitation. The gains in prior specification flexibility are illustrated via simulation studies and an application to a dataset concerning the spatiotemporal evolution of PM10 measurements in Germany.
