A spatiotemporal ensemble model to predict gross beta particulate radioactivity across the contiguous United States
Li, Longxiang ; Blomberg, Annelise J. LU
; Lawrence, Joy
; Réquia, Weeberb J.
; Wei, Yaguang
; Liu, Man
; Peralta, Adjani A.
and Koutrakis, Petros
(2021)
In Environment International
156.
- Abstract
Particulate radioactivity, a characteristic of particulate matter, is primarily determined by the abundance of radionuclides that are bound to airborne particulates. Exposure to high levels of particulate radioactivity has been associated with negative health outcomes. However, there are currently no spatially and temporally resolved particulate radioactivity data for exposure assessment purposes. We estimated the monthly distributions of gross beta particulate radioactivity across the contiguous United States from 2001 to 2017 with a spatial resolution of 32 km, via a multi-stage ensemble-based model. Particulate radioactivity was measured at 129 RadNet monitors across the contiguous U.S. In stage one, we built 264 base learning models... (More)
Particulate radioactivity, a characteristic of particulate matter, is primarily determined by the abundance of radionuclides that are bound to airborne particulates. Exposure to high levels of particulate radioactivity has been associated with negative health outcomes. However, there are currently no spatially and temporally resolved particulate radioactivity data for exposure assessment purposes. We estimated the monthly distributions of gross beta particulate radioactivity across the contiguous United States from 2001 to 2017 with a spatial resolution of 32 km, via a multi-stage ensemble-based model. Particulate radioactivity was measured at 129 RadNet monitors across the contiguous U.S. In stage one, we built 264 base learning models using six methods, then selected nine base models that provide different predictions. In stage two, we used a non-negative geographically and temporally weighted regression method to aggregate the selected base learner predictions based on their local performance. The results of block cross-validation analysis suggested that the non-negative geographically and temporally weighted regression ensemble learning model outperformed all base learning model with the smallest rooted mean square error (0.094 mBq/m3). Our model provided an accurate estimation of particulate radioactivity, thus can be used in future health studies.
(Less)
- author
- Li, Longxiang
; Blomberg, Annelise J.
LU
; Lawrence, Joy
; Réquia, Weeberb J.
; Wei, Yaguang
; Liu, Man
; Peralta, Adjani A.
and Koutrakis, Petros
- organization
- publishing date
- 2021
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Geographically and temporally weighted regression, Particulate radioactivity, Spatiotemporal ensemble learning, Statistical learning
- in
- Environment International
- volume
- 156
- article number
- 106643
- publisher
- Elsevier
- external identifiers
-
- scopus:85110423911
- pmid:34020300
- ISSN
- 0160-4120
- DOI
- 10.1016/j.envint.2021.106643
- language
- English
- LU publication?
- yes
- id
- f3688726-e6bb-4f32-9785-f65074ff5737
- date added to LUP
- 2021-08-19 14:58:15
- date last changed
- 2024-07-13 17:09:39
@article{f3688726-e6bb-4f32-9785-f65074ff5737,
abstract = {{<p>Particulate radioactivity, a characteristic of particulate matter, is primarily determined by the abundance of radionuclides that are bound to airborne particulates. Exposure to high levels of particulate radioactivity has been associated with negative health outcomes. However, there are currently no spatially and temporally resolved particulate radioactivity data for exposure assessment purposes. We estimated the monthly distributions of gross beta particulate radioactivity across the contiguous United States from 2001 to 2017 with a spatial resolution of 32 km, via a multi-stage ensemble-based model. Particulate radioactivity was measured at 129 RadNet monitors across the contiguous U.S. In stage one, we built 264 base learning models using six methods, then selected nine base models that provide different predictions. In stage two, we used a non-negative geographically and temporally weighted regression method to aggregate the selected base learner predictions based on their local performance. The results of block cross-validation analysis suggested that the non-negative geographically and temporally weighted regression ensemble learning model outperformed all base learning model with the smallest rooted mean square error (0.094 mBq/m<sup>3</sup>). Our model provided an accurate estimation of particulate radioactivity, thus can be used in future health studies.</p>}},
author = {{Li, Longxiang and Blomberg, Annelise J. and Lawrence, Joy and Réquia, Weeberb J. and Wei, Yaguang and Liu, Man and Peralta, Adjani A. and Koutrakis, Petros}},
issn = {{0160-4120}},
keywords = {{Geographically and temporally weighted regression; Particulate radioactivity; Spatiotemporal ensemble learning; Statistical learning}},
language = {{eng}},
publisher = {{Elsevier}},
series = {{Environment International}},
title = {{A spatiotemporal ensemble model to predict gross beta particulate radioactivity across the contiguous United States}},
url = {{http://dx.doi.org/10.1016/j.envint.2021.106643}},
doi = {{10.1016/j.envint.2021.106643}},
volume = {{156}},
year = {{2021}},
}