Wednesday, May 21, 2025

Time: 16:10-17:00

venue: Mathematics Building Room 527

Speaker:  Prof. Prof. Hong-Ding Yang (Department of Applied Mathematics, National Chiayi University)

Title: A Bayesian Quantile Hierarchical Model for Spatiotemporal Risk Assessment of Zero-Inflated Count Data

Abstract:

This study develops a spatiotemporal modeling framework for analyzing zero-inflated ecstasy incident data in Kaohsiung City, Taiwan. We employ a Bayesian hierarchical model to account for spatial and temporal correlations, incorporating variable selection based on the Deviance Information Criterion (DIC). Quantile regression is embedded into the model structure to capture heterogeneity across the distribution. Inference is carried out using Markov Chain Monte Carlo (MCMC) techniques to estimate model parameters and random effects. A novel risk assessment criterion is proposed to identify high-risk areas and periods for ecstasy incidents. The proposed approach is compared with alternative models to evaluate its relative performance. Empirical findings demonstrate that the proposed method enhances understanding of spatiotemporal dynamics in ecstasy-related offenses and provides a practical decision-making tool for public safety authorities.

Keywords: Hierarchical modeling, Markov Chain Monte Carlo, Variable selection, Risk estimation