Prediction and Assessment of Minerals Contamination in Groundwater: Analytical Tools Approach

Abstract

Accumulation of minerals in groundwater over years degrades the water quality and thus affects the surrounding ecosystem if left untreated. Rapid urbanization and industrialization paves way for serious harm to the natural resources; particularly for the water bodies. One such study area is chosen for this analytical investigation to predict the consecutive concentration of important minerals for the next five years with some prediction tools. Artificial Neural Network, Support Vector Machine, and, Deep Learning methods are adopted for prediction analysis. The results of MSE, RMSE, and MAPE in each mentioned method were compared and concluded which performed better for the collected data of mineral concentration. Among these tools, SVM showed better results with less error and efficient accuracy (MSE-64.31, RMSE-8.07, and MAPE-3.92) though the other two techniques gave slight accuracy. The annual rainfall values are highly correlated with the mineral values in which the decreasing trend shows the mineral values when rainfall is higher and vice versa. These predicted values aids in creating awareness among the local residents as well as preventing the further pollution of the groundwater by the accumulation of minerals.