Neuronal Brownian dynamics for salinity of river basins’ water management

Salinization of streams, rivers and other water sources threaten the civilizations, ecologies enduring constituent species that results in rendering the precious water unusable for human chores. Increase in salinity across the flow in streams and wet-lands have been mostly to raise a concern towards salt tolerance to various limits. Hence, it becomes important to monitor the acidity/alkalinity causing water parameters that can be referred to as salinity. The prime measure scale of salinity is the quality of potential-of-hydrogen (pH) present in river waters at two sample locations. Two locations that have been identified by CPCB as per the highly reported pollutants’ level found, have been analysed through artificial-intelligence (AI) conjucted with Multivariate Adaptive Regression Spline (MARS). The hybrid of wavelet neuro-fuzzy inferences with that of MARS (WNF-MARS) predicted with more accuracy. Simulation of performance measures: root meant square error (RMSE); mean absolute error (MAE); goodness-of-fit (R2) together with their execution time for the three prototypes provided remarkable results. RMSE outcomes diminish on the whole on applying the training and validating data division in Wavelet conjucted MARS and WNF-MARS as compared to studying the data through MARS. Goodness-of-fit statistic analysed the concentration levels of salinity in the river at the identified sites. Thus, it is observed from this study that the pH levels provide future estimation of inapt quality of water at the source, so that it prohibits the further-decay of water consumed in the ecosystem. Thus, these predictors would be helpful towards formulation of strategies for protection of vegetation and other required purposes.