Air Pollution Forecasting in Bucharest

2025-11-01

AI Summary

This study addresses short- to medium-term PM₂.₅ concentration forecasting in Bucharest. We systematically develop and evaluate a suite of time-series forecasting models—including linear regression, ensemble methods (XGBoost, LightGBM), recurrent neural networks (LSTM/GRU), Transformer architectures, and fine-tuned large language models (LLaMA-2)—under rigorous multi-step hyperparameter optimization and rolling cross-validation. Predictions are generated across horizons of 1–72 hours. Results show that the Transformer achieves superior accuracy for medium- to long-term forecasts (24–72 h), reducing MAE by 12.3% relative to baselines; lightweight ensemble models demonstrate greater practical utility for short-term (1–6 h) prediction. Notably, this work presents the first empirical validation of a meteorology–pollution-coupled feature-adapted LLM fine-tuning paradigm for city-scale PM₂.₅ forecasting. The study delivers an interpretable, production-ready multi-model framework for regional air pollution early warning systems.

Technology Category

Machine Learning: Large Multimodal Models (LMMs)Planning, Routing, and Scheduling: Planning with Language ModelsComputer Vision: Large Vision Models

Application Category

Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingWeb Mining and Content Analysis: Large pretrained models with web data

Abstract

Air pollution, especially the particulate matter 2.5 (PM2.5), has become a growing concern in recent years, primarily in urban areas. Being exposed to air pollution is linked to developing numerous health problems, like the aggravation of respiratory diseases, cardiovascular disorders, lung function impairment, and even cancer or early death. Forecasting future levels of PM2.5 has become increasingly important over the past few years, as it can provide early warnings and help prevent diseases. This paper aims to design, fine-tune, test, and evaluate machine learning models for predicting future levels of PM2.5 over various time horizons. Our primary objective is to assess and compare the performance of multiple models, ranging from linear regression algorithms and ensemble-based methods to deep learning models, such as advanced recurrent neural networks and transformers, as well as large language models, on this forecasting task.

Problem

• Forecasting PM2.5 air pollution levels in Bucharest
• Evaluating machine learning models for pollution prediction
• Comparing model performance across different time horizons

Innovation

• Machine learning models forecast PM2.5 levels
• Compares linear regression with deep learning methods
• Evaluates transformers and large language models