PREDICTING AND ANALYZING WATER QUALITY USING MACHINE LEARNING BASED MODEL: A CASE STUDY FOR KANCHEEPURAM WATERSHED

The deteriorating quality of ground and surfaces water resources is one of the burning and most troublesome issues in front of bio organisms. The impacts of contaminated groundwater are extensive, effecting life in every aspect. Hence, management of ground and surfaces water are vital in order to optimize the groundwater as well as surfaces water quality. The impacts of polluted groundwater can be attempted economically if large data set are analysed and groundwater quality is forecasted ahead of time. Such problem has been raise by many researcher in previous research, still, more research work needs to improvement, efficiency, consistency, accurateness as well as utility of the current water quality assessment techniques. In present research work, the performance of artificial intelligence technique including artificial neural network (ANN) for predicting water quality components of Kancheepuram Watershed located in the extreme south of India. The objective of present research work is to generate a water quality forecast model with the help of known water quality factors using machine learning based ANN model and time-series data analysis. Under research objective uses the water quality historical data of year 1985 to 2018, with post and monsoon time interval. Data is obtained from the Chennai metro water board, Tamil Nadu, India. For present research work, total 11 ground water quality parameters (Cl, EC, HCO3, SO4, NO3, pH, Ca, K, Mg, and Na) are taken which influence ground water quality. Comparison of outcomes of ANN model with observed water quality values shows that although this model has acceptable performance for predicting the components of water quality, ANN accuracy is good. For model performance evaluating using statistical tests like MSE, RMSE, DDR Index and Regression Analysis. According to the statistical tests declared that ANN is good predictor model.