Statitical Modelling of highly schewed soil PCB data with many zeros

Authors: Eunjung Lim*, University of Maryland
Topics: Soils, Spatial Analysis & Modeling, Medical and Health Geography
Keywords: Soil contamination, Skewed data, Spatial Regression, Mixture Model
Session Type: Paper
Day: 4/6/2019
Start / End Time: 5:00 PM / 6:40 PM
Room: Truman, Marriott, Mezzanine Level
Presentation File: No File Uploaded


Polychlorinated Biphenyls (PCBs) are harmful toxic chemicals to human and environment. PCBs were produced in Anniston, AL from 1929 until 1971. Although the commerce of PCBs prohibited since 1976, Landfills of PCBs still remain as contamination sources. Modelling of PCB contamination levels in soils can be an indicator of quantitative estimate of human PCB exposure and health effect. Usually, the first step of modelling is analyzing data distribution of soil PCB levels. Soil PCB contamination data are characterized by highly skewed data inflating at zero. This research aims to describe and predict the effects of covariates on PCB contamination in soil and apply and compare different modeling approaches for skewed data with many zeros.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login