Computer Science > Machine Learning
[Submitted on 26 Feb 2025]
Title:A HEART for the environment: Transformer-Based Spatiotemporal Modeling for Air Quality Prediction
View PDF HTML (experimental)Abstract:Accurate and reliable air pollution forecasting is crucial for effective environmental management and policy-making. llull-environment is a sophisticated and scalable forecasting system for air pollution, inspired by previous models currently operational in Madrid and Valladolid (Spain). It contains (among other key components) an encoder-decoder convolutional neural network to forecast mean pollution levels for four key pollutants (NO$_2$, O$_3$, PM$_{10}$, PM$_{2.5}$) using historical data, external forecasts, and other contextual features. This paper investigates the augmentation of this neural network with an attention mechanism to improve predictive accuracy. The proposed attention mechanism pre-processes tensors containing the input features before passing them to the existing mean forecasting model. The resulting model is a combination of several architectures and ideas and can be described as a "Hybrid Enhanced Autoregressive Transformer", or HEART. The effectiveness of the approach is evaluated by comparing the mean square error (MSE) across different attention layouts against the system without such a mechanism. We observe a significant reduction in MSE of up to 22%, with an average of 7.5% across tested cities and pollutants. The performance of a given attention mechanism turns out to depend on the pollutant, highlighting the differences in their creation and dissipation processes. Our findings are not restricted to optimizing air quality prediction models, but are applicable generally to (fixed length) time series forecasting.
Submission history
From: Norbert Bodendorfer [view email][v1] Wed, 26 Feb 2025 10:54:27 UTC (36 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.