In this article, we present new results of experiments comparing the effectiveness of regression methods in their ability to predict the NSW classification of the level of PM2.5 particles in the air. We used an extensive data set for Wagga Wagga obtained from the DustWatch program and the Bureau of Meteorology, Australia over a twelve months period. The best outcomes were obtained using Additive Regression method based on Isotonic Regression, which is a novel iterative method.

machine learning, multilabel classification, regression, air pollution.