Water Quality Analysis and Prediction of River Water using Support Vector Machine Model

Water quality analysis is an important part of the real-time water monitoring system. Ensures water quality is maintained following the Bureau of Indian Standards (BIS). It is a necessary step to control any pollution that could adversely affect the quality of river water and the river ecosystem. Therefore, this paper is an attempt to develop a support vector machine (SVM) model for predicting the water quality of the Krishna River. The analysis of the data is based on the parameters like temperature, potential hydrogen (pH), dissolved oxygen (DO), Nitrate (NO3), electrical conductivity (EC), total dissolved solids (TDS), biological oxygen demand (BOD), and total dissolved solids (TDS). The proposed model can be considered an effective means to map the status of water quality in the river basin. The analysis is performed by comparing the two kernel functions namely sigmoid, and polynomial, of the SVM technique. The intention here is to measure the efficiency and adequacy of these functions in water monitoring applications. The SVM regression using the ten-fold cross-validation method is applied to the sample data to find the exact drift between the predicted values and the actual value. The root mean square error (RMSE) for the three seasons are analyzed using the different kernel functions and the polynomial function with the efficiency of 97.14 % is found to the suited kernel function for water quality monitoring compared to sigmoid kernel function with the efficiency of 51.43%.