Properties of Minimum Contrast Estimator for Multivariate Spatial Point Processes

• Activity Number: 59 - Invited E-Poster Session I
• Type: Invited
• Date/Time: Sunday, August 8, 2021 : 5:45 PM to 6:30 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Mikyoung Jun* and Lin Zhu and Junho Yang and Scott Cook and Wenlin Dai
• Companies: University of Houston and Texas A&M University and Texas A&M University and Texas A&M University and Renmin University
• Keywords: Minimum contrast; spatial point patterns; terrirosm data; Multivariate point processes

Abstract

Minimum contrast (MC) estimation method for spatial and spatial point process models are commonly used as a tool to get initial values for more computationally intensive likelihood based method or MCMC. It is also a popular tool to get final parameter estimates, when the computation based on likelihood or MCMC procedures is intensive. However, not much has been studied for properties of minimum contrast estimators for spatial point processes. For multivariate case, MC method has not been utilized much at all. In this work, we propose a procedure for MC estimation for multivariate spatial point processes and present results on estimators' properties. Thorough simulation results will be presented. We apply proposed method to modeling bivariate spatial point pattern by terrorist attack patterns in Nigeria. This is a joint work with Junho Yang, Lin Zhu, Wenlin Dai, and Scott Cook.