Open Access

ARTICLE

Accurate Multi-Site Daily-Ahead Multi-Step PM_2.5 Concentrations Forecasting Using Space-Shared CNN-LSTM

Xiaorui Shao, Chang Soo Kim^*

Department of Information Systems, Pukyong National University, Busan, 608737, Korea

* Corresponding Author:Chang Soo Kim. Email:

Computers, Materials & Continua 2022, 70(3), 5143-5160. https://doi.org/10.32604/cmc.2022.020689

Received 03 June 2021; Accepted 05 August 2021; Issue published 11 October 2021

Abstract

Accurate multi-step PM_2.5 (particulate matter with diameters ≤2.5um) concentration prediction is critical for humankinds’ health and air population management because it could provide strong evidence for decision-making. However, it is very challenging due to its randomness and variability. This paper proposed a novel method based on convolutional neural network (CNN) and long-short-term memory (LSTM) with a space-shared mechanism, named space-shared CNN-LSTM (SCNN-LSTM) for multi-site daily-ahead multi-step PM_2.5 forecasting with self-historical series. The proposed SCNN-LSTM contains multi-channel inputs, each channel corresponding to one-site historical PM_2.5 concentration series. In which, CNN and LSTM are used to extract each site's rich hidden feature representations in a stack mode. Especially, CNN is to extract the hidden short-time gap PM_2.5 concentration patterns; LSTM is to mine the hidden features with long-time dependency. Each channel extracted features are merged as the comprehensive features for future multi-step PM_2.5 concentration forecasting. Besides, the space-shared mechanism is implemented by multi-loss functions to achieve space information sharing. Therefore, the final features are the fusion of short-time gap, long-time dependency, and space information, which enables forecasting more accurately. To validate the proposed method's effectiveness, the authors designed, trained, and compared it with various leading methods in terms of RMSE, MAE, MAPE, and R² on four real-word PM_2.5 data sets in Seoul, South Korea. The massive experiments proved that the proposed method could accurately forecast multi-site multi-step PM_2.5 concentration only using self-historical PM_2.5 concentration time series and running once. Specifically, the proposed method obtained averaged RMSE of 8.05, MAE of 5.04, MAPE of 23.96%, and R² of 0.7 for four-site daily ahead 10-hour PM_2.5 concentration forecasting.

---

Keywords

---