Title: Water quality index prediction using artificial neural network: a case study of Selangor River, Malaysia

Authors: Jia Jun Tan; Senthil Kumar Arumugasamy; Fang Yenn Teo

Addresses: Department of Civil Engineering, University of Nottingham Malaysia, Semenyih, 43500 Selangor, Malaysia ' Department of Chemical and Environmental Engineering, University of Nottingham Malaysia, Semenyih, 43500 Selangor, Malaysia ' Department of Civil Engineering, University of Nottingham Malaysia, Semenyih, 43500 Selangor, Malaysia

Abstract: Rapid urban development often leads to deterioration of river water quality, and water quality index (WQI) is a number that represents the water quality of a water body. According to Department of Environment, Malaysia parameters used to calculate WQI are dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), potential of hydrogen (pH), suspended solids (SS) and ammoniacal nitrogen (AN). Data was collected from the ASEAN Working Group on Water Resources Management website in Sungai-Selangor section. Two ANN models were developed; a prediction model to predict the current WQI, and a forecasting model to predict the future WQI. The prediction model gave good results with very low overall root mean squared error (1.15), an excellent overall regression value (0.97874), and a high correlation with the actual WQI (99.94%). The forecasting model did not provide good result with the overall RMSE of 4.80 and overall regression value of 0.752.

Keywords: artificial neural networks; ANNs; water quality index; WQI; dissolved oxygen; DO; biochemical oxygen demand; BOD; chemical oxygen demand; COD; potential of hydrogen; pH; suspended solids; SS; ammoniacal nitrogen; AN; Malaysia.

DOI: 10.1504/IJSAMI.2025.143101

International Journal of Sustainable Agricultural Management and Informatics, 2025 Vol.11 No.1, pp.48 - 71

Received: 21 Jun 2023
Accepted: 19 Sep 2023
Published online: 03 Dec 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article