Estimation of Residential Radon Concentration in Pennsylvania Counties by Data Fusion

TL;DR

This paper introduces a data fusion approach using a multi-sample density ratio model to estimate residential radon levels across Pennsylvania counties, improving regional risk assessment accuracy.

Contribution

It presents a novel data fusion method that combines multiple data sources and periods for more precise radon concentration estimation in specific counties.

Findings

Effective estimation of radon distribution in Beaver County and neighbors.

Improved confidence intervals for radon threshold probabilities.

Method applicable to other regional environmental data analyses.

Abstract

A data fusion method for the estimation of residential radon level distribution in any Pennsylvania county is proposed. The method is based on a multi-sample density ratio model with variable tilts and is applied to combined radon data from a reference county of interest and its neighboring counties. Beaver county and its four immediate neighbors are taken as a case in point. The distribution of radon concentration is estimated in each of six periods, and then the analysis is repeated combining the data from all the periods to obtain estimates of Beaver threshold probabilities and the corresponding confidence intervals.