Statistics Seminar
By Guanjie Lyu
Thusrday, November 7, 2024, at 3:00 pm
University of Windsor | In-person | Lambton Tower 9-118
Title: Semiparametric copula-based quantile regression for semicontinuous data
Abstract: This study introduces a copula-based two-part quantile regression model tailored for analyzing semicontinuous data, which contain a mixture of zeros and positive continuous values. The model separates the occurrence and intensity components, allowing for more flexible and accurate estimation of complex dependence structures in healthcare outcome. Asymptotic properties are established and the performance is investigated through simulations. Applied to healthcare data, the proposed model offers nuanced insights into the relationship between social deprivation and uncompensated care burdens, outperforming competing approaches in capturing nonlinear dependencies and distributional features.
Counts toward seminar attendance for MSc and PhD students in Math & Stats