UMD     This Site

ISR-affiliated Professor Benjamin Kedem (Mathematics), his Ph.D. student Xuze Zhang, and Saumyadipta Pyne of the University of Pittsburgh School of Public Health have proposed a data fusion method for estimating residential radon level distribution in geographic areas where exposure to underground radon may be an issue.

Radon (specifically Radon-222), is a tasteless, colorless and odorless radioactive gas. It is a product of Uranium-238 and Radium-226, both of which are naturally abundant in the soil, particularly in certain geologic regions. Radon is known worldwide as a carcinogen, as its inhaled decay products can get trapped in the lungs and induce DNA mutations. Exposure to radon is the leading cause of lung cancer among non-smokers. In regions with high concentrations of radon, it is both an occupational hazard to miners and a serious risk factor for the general population, who are exposed to the gas in their residences via cracks in basement floors, walls, and so on. In these regions, homeowners are encouraged to outfit their basements with radon detectors and sometimes special mitigating ventilation systems. The presence of radon can affect real estate values.

The paper Estimation of Residential Radon Concentration in Pennsylvania Counties by Data Fusion shows how to estimate the radon concentration distribution in a county of interest by combining or fusing the county’s radon data with data from its neighboring counties, using the so-called density ratio model (DRM) with an extension of variable tilts. Such flexibility is a new feature that mitigates the problem of misspecified DRM.

In the paper, Beaver County in western Pennsylvania, 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 threshold probabilities and the corresponding confidence intervals.

The authors follow a two-stage procedure. First, they estimate the radon concentration distribution for each reference county for each period by DRM, using combined data from the reference county and its neighbors. If the radon distributions in each of the six periods behave similarly, the authors proceed next to combine the data from the different periods and use the DRM again to get an improved estimate of the radon concentration distribution for the reference county. If the radon distributions in each of the six periods behave differently, the results from the last period only are used.

The results are useful for public health policy and real estate property appraisal.

Related Articles:
Computational framework automatically optimizes the shape of tissue engineered vascular grafts
New model for U.S. liver transplant allocation could reduce disparities, save lives
A combined semantic and machine learning approach for smart city digital twins
Public health planners: Free resources for emergency health clinics
New model predictive control framework improves reactive navigation for autonomous robots
New simulator models 3D DRAM thermal characteristics and timings
Port-Hamiltonian modeling formalism framework proposed for swarms
A novel statistical idea: 'Down-Up' sequences that 'capture' small tail probabilities
Brain and Behavior Initiative Hosts 3rd Annual Seed Grant Symposium
Fu, Ryzhov, Qu are issued patent for B2B optimal bidding

January 8, 2020

«Previous Story  



Current Headlines

Engineering Robotic Starfish

UMD receives two new DOE Building Technologies awards

Pines: Stand in Solidarity, Unite Against Injustice

Promoting Diversity and Addressing Unconscious Bias

Alumnus Hyun Jung and colleagues claim top spot in MoNuSAC challenge

Pines, UMD Engineers Help Prince George’s County Reopening Efforts

Data visualization aids the public's pandemic understanding

Ingestible device research advances, enters new phase

Shneiderman's human-computer interaction advocacy profiled in The New York Times

Behtash Babadi promoted to associate professor

Back to top  
Home Clark School Home UMD Home