Hamed Karimian1,2, Qi Li 2, Chunlin Wu2, Yanlin Qi2, Yuqin Mo2, Gong Chen2,4, Xianfeng Zhang2, Sonali Sachdeva3


School of Architecture, Surveying and Mapping Engineering, Jiangxi University of Science and Technology, Jiangxi 341000, China
School of Earth and Space Science, Peking University, Beijing 100871, China
Kavli Institute for Astronomy and Astrophysics, Peking University, Beijing 100871, China
Laboratory of Ocean Environmental Big Data Analysis and Application, Shenzhen 518055, China



Received: January 12, 2019
Revised: April 1, 2019
Accepted: May 10, 2019
Download Citation: ||https://doi.org/10.4209/aaqr.2018.12.0450 

  • Download: PDF


Cite this article:
Karimian, H., Li, Q., Wu, C., Qi, Y., Mo, Y., Chen, G., Zhang, X. and Sachdeva, S. (2019). Evaluation of Different Machine Learning Approaches to Forecasting PM2.5 Mass Concentrations. Aerosol Air Qual. Res. 19: 1400-1410. https://doi.org/10.4209/aaqr.2018.12.0450


HIGHLIGHTS

  • Three ML approaches were implemented and evaluated in PM2.5 forecasts.
  • Advanced ML model (without neurons) showed better performance than deep FNN.
  • Proposed hybrid LSTM is robust in forecasting PM2.5 and air pollution levels.
  • Temporal dependencies of data should be considered in air pollution predictions.

ABSTRACT


With the rapid growth in the availability of data and computational technologies, multiple machine learning frameworks have been proposed for forecasting air pollution. However, the feasibility of these complex approaches has seldom been verified in developing countries, which generally suffer from heavy air pollution. To forecast PM2.5 concentrations over different time intervals, we implemented three machine learning approaches: multiple additive regression trees (MART), a deep feedforward neural network (DFNN) and a new hybrid model based on long short-term memory (LSTM). By capturing temporal dependencies in the time series data, the LSTM model achieved the best results, with RMSE = 8.91 µg m–3 and MAE = 6.21 µg m–3. It also explained 80% of the variability (R2 = 0.8) in the PM2.5 concentrations and predicted 75% of the pollution levels, proving that this methodology can be effective for forecasting and controlling air pollution.


Keywords: Air pollution; Machine learning; Neural networks; Deep learning; Prediction.

 



Latest Articles

Impact Factor: 2.735

5-Year Impact Factor: 2.827


SCImago Journal & Country Rank

Enter your email below to receive latest published articles in your field.