Cargando…
Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network
Many low-cost sensors (LCSs) are distributed for air monitoring without any rigorous calibrations. This work applies machine learning with PM(2.5) from Taiwan monitoring stations to conduct in-field corrections on a network of 39 PM(2.5) LCSs from July 2017 to December 2018. Three candidate models w...
Autores principales: | , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2020
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7506620/ https://www.ncbi.nlm.nih.gov/pubmed/32899301 http://dx.doi.org/10.3390/s20175002 |
_version_ | 1783585055468158976 |
---|---|
author | Wang, Wen-Cheng Vincent Lung, Shih-Chun Candice Liu, Chun-Hu |
author_facet | Wang, Wen-Cheng Vincent Lung, Shih-Chun Candice Liu, Chun-Hu |
author_sort | Wang, Wen-Cheng Vincent |
collection | PubMed |
description | Many low-cost sensors (LCSs) are distributed for air monitoring without any rigorous calibrations. This work applies machine learning with PM(2.5) from Taiwan monitoring stations to conduct in-field corrections on a network of 39 PM(2.5) LCSs from July 2017 to December 2018. Three candidate models were evaluated: Multiple linear regression (MLR), support vector regression (SVR), and random forest regression (RFR). The model-corrected PM(2.5) levels were compared with those of GRIMM-calibrated PM(2.5). RFR was superior to MLR and SVR in its correction accuracy and computing efficiency. Compared to SVR, the root mean square errors (RMSEs) of RFR were 35% and 85% lower for the training and validation sets, respectively, and the computational speed was 35 times faster. An RFR with 300 decision trees was chosen as the optimal setting considering both the correction performance and the modeling time. An RFR with a nighttime pattern was established as the optimal correction model, and the RMSEs were 5.9 ± 2.0 μg/m(3), reduced from 18.4 ± 6.5 μg/m(3) before correction. This is the first work to correct LCSs at locations without monitoring stations, validated using laboratory-calibrated data. Similar models could be established in other countries to greatly enhance the usefulness of their PM(2.5) sensor networks. |
format | Online Article Text |
id | pubmed-7506620 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2020 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-75066202020-09-26 Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network Wang, Wen-Cheng Vincent Lung, Shih-Chun Candice Liu, Chun-Hu Sensors (Basel) Article Many low-cost sensors (LCSs) are distributed for air monitoring without any rigorous calibrations. This work applies machine learning with PM(2.5) from Taiwan monitoring stations to conduct in-field corrections on a network of 39 PM(2.5) LCSs from July 2017 to December 2018. Three candidate models were evaluated: Multiple linear regression (MLR), support vector regression (SVR), and random forest regression (RFR). The model-corrected PM(2.5) levels were compared with those of GRIMM-calibrated PM(2.5). RFR was superior to MLR and SVR in its correction accuracy and computing efficiency. Compared to SVR, the root mean square errors (RMSEs) of RFR were 35% and 85% lower for the training and validation sets, respectively, and the computational speed was 35 times faster. An RFR with 300 decision trees was chosen as the optimal setting considering both the correction performance and the modeling time. An RFR with a nighttime pattern was established as the optimal correction model, and the RMSEs were 5.9 ± 2.0 μg/m(3), reduced from 18.4 ± 6.5 μg/m(3) before correction. This is the first work to correct LCSs at locations without monitoring stations, validated using laboratory-calibrated data. Similar models could be established in other countries to greatly enhance the usefulness of their PM(2.5) sensor networks. MDPI 2020-09-03 /pmc/articles/PMC7506620/ /pubmed/32899301 http://dx.doi.org/10.3390/s20175002 Text en © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). |
spellingShingle | Article Wang, Wen-Cheng Vincent Lung, Shih-Chun Candice Liu, Chun-Hu Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title | Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title_full | Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title_fullStr | Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title_full_unstemmed | Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title_short | Application of Machine Learning for the in-Field Correction of a PM(2.5) Low-Cost Sensor Network |
title_sort | application of machine learning for the in-field correction of a pm(2.5) low-cost sensor network |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7506620/ https://www.ncbi.nlm.nih.gov/pubmed/32899301 http://dx.doi.org/10.3390/s20175002 |
work_keys_str_mv | AT wangwenchengvincent applicationofmachinelearningfortheinfieldcorrectionofapm25lowcostsensornetwork AT lungshihchuncandice applicationofmachinelearningfortheinfieldcorrectionofapm25lowcostsensornetwork AT liuchunhu applicationofmachinelearningfortheinfieldcorrectionofapm25lowcostsensornetwork |