Abstract

The situation of surface water in Hoi An city has been facing challenges in terms of pollution levels for more than a decade. In order to estimate three water quality parameters, such as TSS, COD, and BOD, this study introduces the integration of a machine learning model with an adaptive network-based fuzzy inference system (ANFIS), paired with optical and radar remote sensing data. The RMSE, R², and MAE parameters were used to assess the model's performance. With RMSE ranging from 3.52 mg/l to 4.59 mg/l, R² ranging from 0.69-0.82, and MAE ranging from 2.39 mg/l to 3.16 mg/l, the predictive indices produced significant findings. The results show that the concentrations of three high water quality assessment parameters are distributed in the area of ​​Hoi An Ancient Town, Hoai River. This method can be used to assess surface water quality parameters, to quickly monitor the status of the water environment, to provide a method of monitoring water surface quality, which can be the foundation for providing solutions to protect and use clean water sources in Heritage cities