First-order interaction multiple regressions model on water quality index in manjung riverand its tributaries

Abstract

This research demonstrated the procedures in choosing the best model in forecasting the water quality index (WQI) in Manjung River and its tributaries using multiple regressions. Six independent variables which are the water quality parameters and WQI as the dependent variable were included in this data set. The Multiple Regression (MR) models were involved is the first-order interaction with 57 possible models were considered. In this research, the process of getting the best model from the total of 57 possible models had been shown. The backward elimination of variables with the highest p-value was engaged to get the selected model. The best model includes using the first order interaction with variables of (DO, COD, BOD, SS, AN and pH). The best model obtains then been verified by the Mean Absolute Percentage Error (MAPE) calculation to quantify the models' relative overall fit. ©2006-2016 Asian Research Publishing Network (ARPN).