Air Quality Forecasting Using Neural Networks

Cheng Zhao; Mark van Heeswijk; Juha Karhunen

doi:10.1109/SSCI.2016.7850128

Air Quality Forecasting Using Neural Networks

Cheng Zhao, Mark van Heeswijk, Juha Karhunen

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

4 Citations (Scopus)

177 Downloads (Pure)

Abstract

In this paper, a neural network approach is proposed for air quality forecasting based on the air quality time series itself as well as external meteorological records. A regularized version of the Extreme Learning Machine is used as the main model for the forecasts and feature selection is performed to select the most relevant model inputs. The proposed method is evaluated under different approaches for performing spatial data fusion. Experiments show that accuracy is increased by considering meteorological data; that it matters greatly for the model how the spatial aspect of the problem is taken into account; and finally, that the model is generally able to select relevant inputs and provide accurate air quality forecasts.

Original language	English
Title of host publication	2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016
Publisher	IEEE
Number of pages	7
ISBN (Electronic)	9781509042401
DOIs	https://doi.org/10.1109/SSCI.2016.7850128
Publication status	Published - 9 Feb 2017
MoE publication type	A4 Article in a conference publication
Event	IEEE Symposium Series on Computational Intelligence - Athens, Greece Duration: 6 Dec 2016 → 9 Dec 2016

Conference

Conference	IEEE Symposium Series on Computational Intelligence
Abbreviated title	SSCI
Country	Greece
City	Athens
Period	06/12/2016 → 09/12/2016

Access to Document

10.1109/SSCI.2016.7850128

SSCI_AirQualityAccepted author manuscript, 1.18 MB

Fingerprint Dive into the research topics of 'Air Quality Forecasting Using Neural Networks'. Together they form a unique fingerprint.

Cite this

@inproceedings{96bb8d83bb5a4a2dadcd7e27b067dabc,

title = "Air Quality Forecasting Using Neural Networks",

abstract = "In this paper, a neural network approach is proposed for air quality forecasting based on the air quality time series itself as well as external meteorological records. A regularized version of the Extreme Learning Machine is used as the main model for the forecasts and feature selection is performed to select the most relevant model inputs. The proposed method is evaluated under different approaches for performing spatial data fusion. Experiments show that accuracy is increased by considering meteorological data; that it matters greatly for the model how the spatial aspect of the problem is taken into account; and finally, that the model is generally able to select relevant inputs and provide accurate air quality forecasts.",

author = "Cheng Zhao and {van Heeswijk}, Mark and Juha Karhunen",

year = "2017",

month = feb,

day = "9",

doi = "10.1109/SSCI.2016.7850128",

language = "English",

booktitle = "2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016",

publisher = "IEEE",

address = "United States",

note = "IEEE Symposium Series on Computational Intelligence, SSCI ; Conference date: 06-12-2016 Through 09-12-2016",

}

TY - GEN

T1 - Air Quality Forecasting Using Neural Networks

AU - Zhao, Cheng

AU - van Heeswijk, Mark

AU - Karhunen, Juha

PY - 2017/2/9

Y1 - 2017/2/9

N2 - In this paper, a neural network approach is proposed for air quality forecasting based on the air quality time series itself as well as external meteorological records. A regularized version of the Extreme Learning Machine is used as the main model for the forecasts and feature selection is performed to select the most relevant model inputs. The proposed method is evaluated under different approaches for performing spatial data fusion. Experiments show that accuracy is increased by considering meteorological data; that it matters greatly for the model how the spatial aspect of the problem is taken into account; and finally, that the model is generally able to select relevant inputs and provide accurate air quality forecasts.

AB - In this paper, a neural network approach is proposed for air quality forecasting based on the air quality time series itself as well as external meteorological records. A regularized version of the Extreme Learning Machine is used as the main model for the forecasts and feature selection is performed to select the most relevant model inputs. The proposed method is evaluated under different approaches for performing spatial data fusion. Experiments show that accuracy is increased by considering meteorological data; that it matters greatly for the model how the spatial aspect of the problem is taken into account; and finally, that the model is generally able to select relevant inputs and provide accurate air quality forecasts.

U2 - 10.1109/SSCI.2016.7850128

DO - 10.1109/SSCI.2016.7850128

M3 - Conference contribution

BT - 2016 IEEE Symposium Series on Computational Intelligence, SSCI 2016

PB - IEEE

T2 - IEEE Symposium Series on Computational Intelligence

Y2 - 6 December 2016 through 9 December 2016

ER -