Implementing Support Vector Machine Method for Predicting Urban Area Water Quality

Water quality failure is a long-standing global issue, causing illness, poisoning, spread of diseases, and claiming human lives in urban areas. Urban growth patterns have a substantial impact on stream water quality because they affect pollutant formation, buildup, and wash-off process. To successfully advise urban growth planning and policies, stream water quality must be evaluated and predicted in relation to various urban development patterns. In this work, Support Vector Machine (SVM) is implemented to predict quality of water in urban areas. Initially, data are collected through collected dataset, and the cleaning and normalization techniques are utilized to remove missing values and outliers and normalized the data. Then, SVM method is employed to effectively predict the quality of water. When compared to existing methods like genetic algorithm_extreme gradient boosting (GA_XGBoost), and artificial neural networks (ANNs), implemented SVM achieved high values of 0.953 of $R 2,0.03$ of Root-mean-square error (RMSE), 0.01 of Mean Absolute Error (MAE), 9 of average invalidity percent (AIP), and 99 of average validity percent (AVP) performance result.