Emissions of PM10 from Dolomite Blast from Atte and Ikpeshi communities Mines in Edo State, Nigeria

The problem of dolomite blast PM10 emissions from the mining environment remains a global issue, with Nigeria being notably affected because of the lack of enforceable laws. Consequently, the objective of this paper is to predict the emissions of PM10 from dolomite blast activities from Atte and Ikpeshi communities’ mines in Edo State, Nigeria using machine learning techniques (MLT) which includes artificial neural networks (ANN), gradient boosting regression (GBR), random forest (RF), and multiple linear regression (MLR) models. The results of the statistical analysis showed that the mean values of the input parameters were 2.984 m for stemming, 1.002 m for burden, 1.008 m for spacing, 9.915 m for hole depth, 123.173 kg for charge weight, and 0.254 kg/m³ for powder factor. Meanwhile, the mean value for the output parameter, PM10, was 1591.98 µg/m³. The suitability of the predictive ability of the models ranked ANN as the best model for predicting dolomite blast PM10 emissions with a testing R² value of 0.959. This is as a result of the ANN model's exceptional balance of accuracy and generalization ability.