Abstract

 
Water quality prediction play an essential role in aqua environment management.  The demand for accurate water quality prediction techniques for efficient water resources management. Currently, the Indian pollution control board has set up various monitoring stations to measure water quality frequently. However, the forecast for water quality is currently not being carried out. In this work, machine learning models have been implemented to predict the indices of water quality. The efficiency of logistic Linear regression and AdaBoostRegressor in the prediction of seven major water quality parameters were evaluated. The Tamil Nadu water quality dataset is used in this analysis. The parameters such as pH value, the quantity of oxygen dissolved, total coli form, B.D.O, electric conductivity, the quantity of phosphorus, and nitrate are considered. The assessed error-index value of the applied models showed that the AdaboostRegressor obtains a lesser error-index and it can consider being a more accurate model than the Linear regression model. The entire methodology proposed here is in the context of water quality is based on numerical analysis. While investigating the outcomes of the implemented machine learning models, it is demonstrated that they have nearly over-estimation properties. The proposed models are assessed using the metrics Mean Square Error and R² score the results reflect that AdaboostRegressor predicts the (Water Quality Indices) WQI rate with a Mean Square Error value of 0.8, and R² score rate is 0.41, whereas AdaBoostRegressor with a obtains Mean Square Error (MSE) rate as 0.74 and R² score rate as 0.44.