What question did this study set out to answer?

The aim is to develop a prediction model for PM2.5 concentrations that overcomes existing forecasting challenges.

February 19, 2026Open Access

Time Series-Based PM2.5 Concentration Prediction Model Incorporating Attention Mechanism

Key Points

The aim is to develop a prediction model for PM2.5 concentrations that overcomes existing forecasting challenges.
Introduced OVMD-PeepholeLSTM-attention model for enhanced PM2.5 prediction.
Utilized optimal variational mode decomposition to decompose PM2.5 data into K components.
Applied PeepholeLSTM with an attention mechanism to predict each component.
Conducted evaluations using PM2.5 data from Guangzhou and Shenzhen during 2020-2022.
Achieved a 39% reduction in mean absolute error (MAE) and a 45% reduction in root mean square error (RMSE) in Guangzhou.
Increased the R2 fitting coefficient by 0.0457 in Guangzhou compared to the traditional model.
Improvements in Shenzhen included a 45% reduction for MAE and a 51% reduction for RMSE, with a 0.0765 increase in R2.

Abstract

As a key indicator of air quality, effective forecasting of PM2.5 concentration can provide key technical support for the scientific and precise implementation of air pollution prevention and control. However, predicting PM2.5 concentrations faces challenges such as multiple influencing factors, long-term temporal dependencies, and inherent nonlinearity. Furthermore, traditional Long Short-Term Memory (LSTM) networks not only fail to effectively grasp the dependency relationships in long-time-span data, but also encounter difficulties in fully integrating and exploiting the information of numerous influencing factors. In order to solve these problems, a novel prediction model (OVMD–PeepholeLSTM–attention) for PM2.5 concentration was presented in this study, which includes Peephole Long Short-Term Memory (PeepholeLSTM), optimal variational mode decomposition (OVMD) and an attention mechanism (AM). In this study, K modal components result from the initial decomposition of PM2.5 monitoring data using OVMD. The obtained components are then individually predicted by the PeepholeLSTM–attention model, and the final prediction is reconstructed. The proposed model was comprehensively evaluated on PM2.5 concentration monitoring data sets from Guangzhou and Shenzhen in China from 2020 to 2022, through a series of comparative experiments. The model proposed in this study is shown by experimental results to reduce mean absolute error (MAE) by approximately 39%, root mean square error (RMSE) by 45%, and increases the fitting coefficient (R2) by 0.0457 in Guangzhou compared to the single PeepholeLSTM model. The corresponding improvements in Shenzhen are 45% for MAE, 51% for RMSE, and 0.0765 for R2. This indicates that the model proposed in this paper exhibits higher accuracy in terms of predicting PM2.5 concentrations, and the research results can provide a basis for quantitative assessment and scientific decision-making for the sustainable development of urban ecological environments.

Read Full Paperexternally

Bookmark

View Full Paper

Cite This Study

Cheng et al. (Tue,) studied this question.

synapsesocial.com/papers/6996a8d4ecb39a600b3eff60 https://doi.org/https://doi.org/10.3390/su18042038

Bookmark

View Full Paper