Abstract:With the increase of global economic activities and high energy demand, many countries have raised concerns about air pollution. However, air quality prediction is a challenging issue due to the complex interaction of many factors. In this paper, we propose a multimodal approach for spatio-temporal air quality prediction. Our model learns the multimodal fusion of critical factors to predict future air quality levels. Based on the analyses of data, we also assessed the impacts of critical factors on air quality prediction. We conducted experiments on two real-world air pollution datasets. For Seoul dataset, our method achieved 11% and 8.2% improvement of the mean absolute error in long-term predictions of PM2.5 and PM10, respectively, compared to baselines. Our method also reduced the mean absolute error of PM2.5 predictions by 20% compared to the previous state-of-the-art results on China 1-year dataset.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the long - term air quality prediction in air quality management. Specifically, the authors propose a multimodal method (STAR) for spatio - temporal air quality prediction, especially for the prediction of fine particulate matter (PM 2.5) and coarse particulate matter (PM 10). These particulate matters are of particular concern due to their long - term impact on human health. The paper points out that due to the complex interaction of multiple factors, such as local pollutant emissions, coal - fired power plants, sand - dust activities, seasonal conditions, meteorology, terrain and various human activities, air quality prediction is a challenging problem. In addition, the lack of publicly available data sets containing a large amount of data on key factors is also a major difficulty in this problem. To solve these problems, the STAR model proceeds through the following steps: 1. **Spatial transformation**: Convert the observed data (air quality data and meteorological data) into heat - map images to analyze the spatial variation of air quality. Using the image form enables the model to effectively capture spatial attributes and relationships without explicitly modeling the spatio - temporal dynamics of PM 2.5 or PM 10 and other atmospheric factors. 2. **Feature fusion**: Combine the spatial representation with the time - based features of key factors into an encoded vector. This vector is then concatenated with the feature representation output in the previous time step to form a context vector. Based on this, the model generates heat - map images of two feature maps for predicting particulate matter levels. The paper also conducts experiments to verify the performance of the STAR model on actual air pollution data sets. The experimental results show that the STAR model performs particularly well in long - term prediction. Compared with the baseline method, its long - term prediction mean absolute error for PM 2.5 and PM 10 on the Seoul data set is improved by 11% and 8.2% respectively. On the one - year data set in China, the STAR model also significantly outperforms the previous state - of - the - art methods, with a 20% reduction in the mean absolute error for PM 2.5 prediction. These results demonstrate the effectiveness of the STAR model in handling complex spatio - temporal data and improving the accuracy of air quality prediction.

STAR: Spatio-Temporal Prediction of Air Quality Using A Multimodal Approach

Predicting Air Quality Using a Multi-Scale Spatiotemporal Graph Attention Network

A Model-driven and Data-driven Fusion Framework for Accurate Air Quality Prediction

An Air Pollutants Prediction Method Integrating Numerical Models and Artificial Intelligence Models Targeting the Area around Busan Port in Korea

A Deep Spatial-Temporal Ensemble Model for Air Quality Prediction

Novel Approach for Predicting the Air Quality Index of Megacities through Attention-Enhanced Deep Multitask Spatiotemporal Learning

Deep learning based multimodal urban air quality prediction and traffic analytics

Spatiotemporal prediction of particulate matter concentration based on traffic and meteorological data

A novel method for identifying hotspots and forecasting air quality through an adaptive utilization of spatio-temporal information of multiple factors

Real-time Air Pollution prediction model based on Spatiotemporal Big data

A Multi-Modal Deep-Learning Air Quality Prediction Method Based on Multi-Station Time-Series Data and Remote-Sensing Images: Case Study of Beijing and Tianjin

Modeling the Air Pollution Process Using a Novel Multi-Site and Multi-Scale Method with Adaptive Utilization of Spatio-Temporal Information.

Predicting air quality via multimodal AI and satellite imagery

Human-model hybrid Korean air quality forecasting system

Multi-Site and Multi-Pollutant Air Quality Data Modeling

Air quality index prediction via multi-task machine learning technique: spatial analysis for human capital and intensive air quality monitoring stations

A novel framework for high resolution air quality index prediction with interpretable artificial intelligence and uncertainties estimation

Temporal-Spatial Aggregated Urban Air Quality Inference With Heterogeneous Big Data

Daily PM2.5 concentration prediction based on variational modal decomposition and deep learning for multi-site temporal and spatial fusion of meteorological factors

Air Quality Prediction Model Based on Spatiotemporal Data Analysis and Metalearning

A Daily Air Pollutant Concentration Prediction Framework Combining Successive Variational Mode Decomposition and Bidirectional Long Short-Term Memory Network