Prediction accuracy of ordinary least square and artificial neural network (mlpregressor) in prediction of rainfall and ore productions

AfuJ, Johnson Damilola and Adebayo, Babatunde and Obabiyi, Yusuf Kayode and Ekun, Adekunle Agbhaseh (2025) Prediction accuracy of ordinary least square and artificial neural network (mlpregressor) in prediction of rainfall and ore productions. Open Access Research Journal of Engineering and Technology, 9 (1). 021-028. ISSN 2783-0128

Application of ordinary least square (OLS) and artificial neural network (ANN) in prediction of rainfall and production (ore and waste) for better prediction accuracy and sustainability of mining production is hereby study in this paper. The degree and accuracy of both OLS and ANN was examined using rainfall and production data over period of 5 years in an open pit mine in South Eastern part of Nigeria, Ebonyi State. The OLS model performed best compared with MLPRegressor in predicting the project cumulative total using the rainfall days’ values. The OLS could be used to predict 19% of the variance in the project cumulative total. Using daily handling waste, daily handling ore and rainfall days as input, the OLS model predicted 93% variance in the project cumulative total, while the MLP Regressor performed best as it can be used to predict 99% variance in the Project Cummulative total. Hence, MLP Regressor is a better tool in future prediction when compared with OLS.