A team of researchers from Beijing Technology and Business University (BTBU) and the Institute of Environmental Assessment and Water Research (IDAEA-CSIC), in collaboration with European partners, has unveiled LPO-XGBoost—a cutting-edge machine learning model designed to enhance the identification of PM10 pollution sources in urban environments.

By integrating traditional Positive Matrix Factorization (PMF) with a nonlinear XGBoost algorithm, the model offers a significant leap forward in air quality analysis. Tested across 21 European monitoring sites, LPO-XGBoost demonstrated high predictive accuracy (R² = 0.88), particularly excelling at pinpointing sources such as sea salt and biomass burning.

This innovation is part of the RI-URBANS initiative, which aims to bridge the gap between scientific advances and practical air quality management across European cities.

Dig more into this interesting study led by Ying Liu here.