Air Pollution Forecasting Using a Deep Learning Model Based on 1D Convnets and Bidirectional GRU

Qing, Tao; Liu, Fang; Li, Yong; Sidorov, Denis

doi:10.1109/access.2019.2921578

Cited by 251 publications

(111 citation statements)

References 12 publications

Supporting

Mentioning

105

Contrasting

Unclassified

Order By: Relevance

“…For the air pollution domain, the metrological data inputs have a great influence on the variability of air pollutants. The increasing number of input data does not guarantee enhancement of the accuracy of the prediction model, in [52]. It is highly recommended those inputs that are highly correlated with the predicted value (air pollutants concentrations) be included.…”

Section: Discussion About Related Workmentioning

confidence: 99%

“…Li, at al., proposed a novel long shortterm memory neural network extended (LSTME) model that inherently considers spatiotemporal correlations for air pollutant concentration prediction [24]. TAO, et al, proposed a model that combines RNN and CNN for air pollution forecasting [52]. Sun, et al, proposed a spatial temporal PM2.5 concentration prediction framework using GRU, which is an extension of RNN [53].…”

Section: Recurrent Neural Network For Forecasting Air Pollutionmentioning

confidence: 99%

“…In most machine learning applications, feature importance is an essential step which can be performed in numerous ways. For timeforecasting prediction problem, most of the previous works applied mathematical correlation to find a relationship between input and output variables [51,24,52,53,56]. When there are many features to be entered into the network for training, then finding the correlation between the target output value and these features will reduce complexity of training and enhance performance.…”

Section: Ensemble Models Selectionmentioning

confidence: 99%

See 2 more Smart Citations

On the Ensemble of Recurrent Neural Network for Air Pollution Forecasting: Issues and Challenges

Surakhi¹,

Serhan²,

Salah³

2020

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

Time-series is a sequence of observations that are taken sequentially over time. Modelling a system that generates a future value from past observations is considered as time-series forecasting system. Recurrent neural network is a machine learning method that is widely used in the prediction of future values. Due to variant improvements on recurrent neural networks, choosing of the best model for better prediction generation is dependent on problem domain and model design characteristics. Ensemble forecasting is more accurate than single model due to the combination of more than one model for forecasting. Designing an ensemble model of recurrent neural network for time-series forecasting applications would enhance prediction accuracy and improve performance. This paper highlights some of the challenges that are faced by the design of the ensemble model of different recurrent neural network versions, and surveys some of the most relevant works in order to give a direction of how to conduct ensemble learning research in the future. Based on the reviewed literature, we propose a framework for time-series forecasting based on the using of ensemble technique.

show abstract

Section: Discussion About Related Workmentioning

confidence: 99%

Section: Recurrent Neural Network For Forecasting Air Pollutionmentioning

confidence: 99%

Section: Ensemble Models Selectionmentioning

confidence: 99%

See 1 more Smart Citation

On the Ensemble of Recurrent Neural Network for Air Pollution Forecasting: Issues and Challenges

Surakhi¹,

Serhan²,

Salah³

2020

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

show abstract

“…There are two clusters for wind speed forecasting model and four clusters for wind power forecasting. Therefore, the final outputs for wind power and wind speed, respectively, can be expressed as follows: y wind power = w wp1 y wind power1 + w wp2 y wind power2 + w wp3 y wind power3 (12) y wind speed = w ws1 y wind speed1 + w ws2 y wind speed2 + w ws3 y wind speed3…”

Section: Parametersmentioning

confidence: 99%

Wind Speed and Power Ultra Short-Term Robust Forecasting Based on Takagi–Sugeno Fuzzy Model

Liu

Dreglea

2019

Energies

Self Cite

View full text Add to dashboard Cite

Accurate wind power and wind speed forecasting remains a critical challenge in wind power systems management. This paper proposes an ultra short-time forecasting method based on the Takagi–Sugeno (T–S) fuzzy model for wind power and wind speed. The model does not rely on a large amount of historical data and can obtain accurate forecasting results though efficient linearization. The proposed method employs meteorological measurements as input. Next, the antecedent and the consequent parameters of the forecasting model are identified by the fuzzy c-means clustering algorithm and the recursive least squares method. From these components, the T–S fuzzy model is obtained. Wind farms located in China (Shanxi Province) and in Ireland (County Kerry) are considered as cases with which to validate the proposed forecasting method. The forecasting results are compared with results from the contemporary machine learning-based models including support vector machine (SVM), the combined model of SVM and empirical mode decomposition, and back propagation neural network methods. The results show that the proposed T–S fuzzy model can effectively improve the precision of the short-term wind power forecasting.

show abstract

“…Recently, intelligent and data-driven methods have become increasingly famous for the prediction of air pollution [33][34][35][36][37]. Among them, machine learning methods have been reported to deliver higher performance in terms of accuracy, robustness, and lower computational power in dealing with uncertainties and big data [38][39][40][41].…”

Section: Introductionmentioning

confidence: 99%

Developing an ANFIS-PSO Model to Predict Mercury Emissions in Combustion Flue Gases

et al. 2019

View full text Add to dashboard Cite

Accurate prediction of mercury content emitted from fossil-fueled power stations is of the utmost importance for environmental pollution assessment and hazard mitigation. In this paper, mercury content in the output gas of power stations’ boilers was predicted using an adaptive neuro-fuzzy inference system (ANFIS) method integrated with particle swarm optimization (PSO). The input parameters of the model included coal characteristics and the operational parameters of the boilers. The dataset was collected from 82 sample points in power plants and employed to educate and examine the proposed model. To evaluate the performance of the proposed hybrid model of the ANFIS-PSO, the statistical meter of MARE% was implemented, which resulted in 0.003266 and 0.013272 for training and testing, respectively. Furthermore, relative errors between the acquired data and predicted values were between −0.25% and 0.1%, which confirm the accuracy of the model to deal non-linearity and represent the dependency of flue gas mercury content into the specifications of coal and the boiler type.

show abstract

Air Pollution Forecasting Using a Deep Learning Model Based on 1D Convnets and Bidirectional GRU

Cited by 251 publications

References 12 publications

On the Ensemble of Recurrent Neural Network for Air Pollution Forecasting: Issues and Challenges

On the Ensemble of Recurrent Neural Network for Air Pollution Forecasting: Issues and Challenges

Wind Speed and Power Ultra Short-Term Robust Forecasting Based on Takagi–Sugeno Fuzzy Model

Developing an ANFIS-PSO Model to Predict Mercury Emissions in Combustion Flue Gases

Contact Info

Product

Resources

About