Data Mining for Extreme Behavior with Application to Ground Level Ozone

• Activity Number: 87
• Type: Invited
• Date/Time: Sunday, August 3, 2014 : 8:30 PM to 10:30 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Brook T. Russell*+ and Daniel S. Cooley and William C. Porter and Colette L. Heald and Brian Reich
• Companies: Colorado State University and Colorado State University and Massachusetts Institute of Technology and Massachusetts Institute of Technology and North Carolina State University
• Keywords: tail dependence ; multivariate regular variation

Abstract

Ground level ozone is a harmful pollutant that can negatively affect people as well as plant species. These negative effects can be intensified when ozone is at extreme levels. In order to better understand the causes of extreme ozone levels, we propose a method to find the linear combination of meteorological covariates that has the highest degree of tail dependence with ozone. This method differs from standard regression methods in that it focuses only on the tail behavior by utilizing the framework of regular variation. Additionally, we propose a measure of tail dependence along with its corresponding estimator which is well-suited for optimization procedures. We present a simulation study which shows that the method can detect complicated conditions leading to extreme responses. We apply the method to ozone data for Atlanta and Charlotte and find similar meteorological drivers for these two Southeastern US cities. To learn the meteorological drivers of extreme ozone we employ an automated model search procedure. Our current work includes a spatial extension of our modeling procedure.