A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction

To avoid the adverse effects of severe air pollution on human health, we need accurate real-time air quality prediction. In this paper, for the purpose of improve prediction accuracy of air pollutant concentration, a deep neural network model with multitask learning (MTL-DBN-DNN), pretrained by a de...

Full description

Saved in:
Bibliographic Details
Main Authors: Jiangeng Li, Xingyang Shao, Rihui Sun
Format: Article
Language:English
Published: Wiley 2019-01-01
Series:Journal of Control Science and Engineering
Online Access:http://dx.doi.org/10.1155/2019/5304535
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832551667056246784
author Jiangeng Li
Xingyang Shao
Rihui Sun
author_facet Jiangeng Li
Xingyang Shao
Rihui Sun
author_sort Jiangeng Li
collection DOAJ
description To avoid the adverse effects of severe air pollution on human health, we need accurate real-time air quality prediction. In this paper, for the purpose of improve prediction accuracy of air pollutant concentration, a deep neural network model with multitask learning (MTL-DBN-DNN), pretrained by a deep belief network (DBN), is proposed for forecasting of nonlinear systems and tested on the forecast of air quality time series. MTL-DBN-DNN model can solve several related prediction tasks at the same time by using shared information contained in the training data of different tasks. In the model, DBN is used to learn feature representations. Each unit in the output layer is connected to only a subset of units in the last hidden layer of DBN. Such connection effectively avoids the problem that fully connected networks need to juggle the learning of each task while being trained, so that the trained networks cannot get optimal prediction accuracy for each task. The sliding window is used to take the recent data to dynamically adjust the parameters of the MTL-DBN-DNN model. The MTL-DBN-DNN model is evaluated with a dataset from Microsoft Research. Comparison with multiple baseline models shows that the proposed MTL-DBN-DNN achieve state-of-art performance on air pollutant concentration forecasting.
format Article
id doaj-art-4a13541e44124d0da6de998dd13239fc
institution Kabale University
issn 1687-5249
1687-5257
language English
publishDate 2019-01-01
publisher Wiley
record_format Article
series Journal of Control Science and Engineering
spelling doaj-art-4a13541e44124d0da6de998dd13239fc2025-02-03T06:00:53ZengWileyJournal of Control Science and Engineering1687-52491687-52572019-01-01201910.1155/2019/53045355304535A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality PredictionJiangeng Li0Xingyang Shao1Rihui Sun2College of Automation, Faculty of Information Technology, Beijing University of Technology, Beijing 100124, ChinaCollege of Automation, Faculty of Information Technology, Beijing University of Technology, Beijing 100124, ChinaCollege of Automation, Faculty of Information Technology, Beijing University of Technology, Beijing 100124, ChinaTo avoid the adverse effects of severe air pollution on human health, we need accurate real-time air quality prediction. In this paper, for the purpose of improve prediction accuracy of air pollutant concentration, a deep neural network model with multitask learning (MTL-DBN-DNN), pretrained by a deep belief network (DBN), is proposed for forecasting of nonlinear systems and tested on the forecast of air quality time series. MTL-DBN-DNN model can solve several related prediction tasks at the same time by using shared information contained in the training data of different tasks. In the model, DBN is used to learn feature representations. Each unit in the output layer is connected to only a subset of units in the last hidden layer of DBN. Such connection effectively avoids the problem that fully connected networks need to juggle the learning of each task while being trained, so that the trained networks cannot get optimal prediction accuracy for each task. The sliding window is used to take the recent data to dynamically adjust the parameters of the MTL-DBN-DNN model. The MTL-DBN-DNN model is evaluated with a dataset from Microsoft Research. Comparison with multiple baseline models shows that the proposed MTL-DBN-DNN achieve state-of-art performance on air pollutant concentration forecasting.http://dx.doi.org/10.1155/2019/5304535
spellingShingle Jiangeng Li
Xingyang Shao
Rihui Sun
A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
Journal of Control Science and Engineering
title A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
title_full A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
title_fullStr A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
title_full_unstemmed A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
title_short A DBN-Based Deep Neural Network Model with Multitask Learning for Online Air Quality Prediction
title_sort dbn based deep neural network model with multitask learning for online air quality prediction
url http://dx.doi.org/10.1155/2019/5304535
work_keys_str_mv AT jiangengli adbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction
AT xingyangshao adbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction
AT rihuisun adbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction
AT jiangengli dbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction
AT xingyangshao dbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction
AT rihuisun dbnbaseddeepneuralnetworkmodelwithmultitasklearningforonlineairqualityprediction