Assessment of Algorithm Performance on Predicting Total Dissolved Solids Using Artificial Neural Network and Multiple Linear Regression for the Groundwater Data

Estimating groundwater quality parameters through conventional methods is time-consuming laboratory measurements for megacities. There a need to develop models that can help decision-makers make policies sustainable reserves. The current study compared the efficiency of multivariate linear regressions (MLR) and artificial neural network (ANN) in prediction total dissolved solids (TDS) three sub-divisions Lahore, Pakistan. data this were collected every quarter year six years. ANN was applied investigate feasibility feedforward, backpropagation networks with training functions T-BR (Bayesian regularization backpropagation), T-LM (Levenberg–Marquardt T-SCG (scaled conjugate backpropagation). Two activation used analyze performance algorithmic functions, i.e., Logsig Tanh. Input pH, electrical conductivity (EC), calcium (Ca2+), magnesium (Mg2+), chloride (Cl−), sulfate (SO42−) predict TDS as an output parameter. computed values by MLR close agreement their respective measured values. Comparative analysis showed root means square error (RMSE) city sub-division Pearson’s coefficient correlation (r) 2.9% 0.981 4.5% 0.978, respectively. Similarly, Farrukhabad sub-division, RMSE r 4.9% 0.952, while 5.5% 0.941, For Shahadra 10.8%, 0.869 ANN, 11.3%, 0.860 MLR. results exhibited model less than Therefore, be employed successfully tool assessment.