Water Quality Analysis Using Artificial Intelligence Conjunction with Wavelet Decomposition

Abstract

Water is life and is the most precious resource on Earth. Earth consists of 70% of water, 2.5% of freshwater, and 1% of easily accessible freshwater; thus, only 0.007% of Earth’s water is accessible. The survival of life on Earth is directly proportional to the presence of water among other important resources. Water remains to be a natural resource with no replacement. In today’s era, where science and technology are growing every hour and innovating new technologies and devices to make life easier and comfortable, but no artificial intelligence could either replicate or replace the need for water on Earth. The present study deals with the qualitative exploration of water quality components like potential of hydrogen (pH), chemical oxygen demand (COD); biochemical oxygen demand (BOD); dissolved oxygen (DO) of Yamuna River at different sample sites. Various sample sites designated for highly reported pollutants using artificial intelligence through least squares support vector regression (LSSVR) and hybrid of wavelet and LSSVR. It is observed that hybrid of wavelet and least squares support vector regression (WLSSVR) predicted good quality accurately among the two prototypes simulated on the basis of the simulation errors which are root–mean-square error (RMSE); mean absolute error (MAE); coefficient of determination (R²); and execution time for both prototypes. RMSE values decrease overall on training and validating via WLSSVR as compared to LSSVR. It is observed that MAE values show a lesser decrease as it is in RMSE; on an average, MAE has lesser variability and R² has a greater variability as per simulations. The simulation is carried out to analyze the level of various pollutants in the Yamuna River at different sites for the consideration of the quality of water. The observed pattern from the study may help for the future prediction of the quality of water parameters, so that it prohibits the further decay of water quality which may prove to be lethal to the environment. These forecasts may be helpful for the formulation of policies, planning, and execution for the protection of the environment and quality of water.