Spatiotemporal Analysis with Transformer Neural Networks: PM2.5 Forecasting
Faculty Mentor
Dr. Weitian Tong
Location
Poster 205
Session Format
Poster Presentation
Academic Unit
Department of Computer Science
Keywords
Allen E. Paulson College of Engineering and Computing Student Research Symposium, Standard Deviation of Error, SDE, Symmetric Mean Absolute Percentage Error, SMAPE, Long-Short Term Memory, LSTM
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Presentation Type and Release Option
Presentation (File Not Available for Download)
Start Date
2022 12:00 AM
January 2022
Spatiotemporal Analysis with Transformer Neural Networks: PM2.5 Forecasting
Poster 205
Air pollution is a potent adversary. Countries still struggle with sparsity of monitoring stations, which creates demand for real-time air quality forecasting. Machine learning algorithms have shown success overcoming challenges of classical prediction model. Currently, these models fail to handle spatiotemporal patterns effectively (McKendry 2002). Inspired by successes in the sequential data processing, we implement and evaluate a Transformer Neural Network as an improved tool for air pollution prediction.
