PREDICTING URBAN SOIL POLLUTION USING MACHINE LEARNING ALGORITHMS

Different cities of the world are facing heavy metal pollution in soils at different levels. Previous studies have found that heavy metal concentrations in urban soils tend to increase with increasing levels of urbanization, indicating a link between heavy metal content in soils and urban expansion. Thus, understanding this relationship and considering factors related to urbanization to create reliable predictions of heavy metal distribution in soils can contribute to effective management of urban health. This study examines the sources, distribution, and environmental effects of heavy metals. These elements accumulate in soil due to vehicle emissions, tire and brake wear, and abrasion of road surfaces, which carry significant environmental and health risks. The presence of heavy metals in road soil can detrimentally affect plant growth, enter the food chain, and pose a direct threat to human health when contaminated soil is ingested, or dust particles are inhaled. In this study, a random forest (RF) machine learning model was applied to predict the extent of heavy metals in soil along highways. The results showed that the RF model has high accuracy in predicting the spatial distribution of heavy metals in soil.