An application of multiple regression for predicting Turbidity of standard water quality for industrial and household consumption

Yusraini Muharni (Department of Industrial Engineering)
Natalia Hartono (Department of Mechanical Engineering, University of Birmingham, Edgbaston, Birmingham)

Journal Industrial Servicess
Vol 7, No 1 (2021): Oktober 2021

A multiple regression approach was applied in this study to predict the Turbidity value of standard water in a water treatment plant. Turbidity is a level of cloudiness in water due to the presence of particles or microorganisms. Turbidity in standard water did not affect human health in terms of hazardous, even though it represents poor quality water. Water treatment plants reduce the cloudiness in water by applying the chlorination process. There are three independent variables of water quality involved to predict turbidity value. They are PH, color-spectrum, and electrical conductivity. The correlation among variables was checked before conducting multiple regression. Color-spectrum has the highest correlation with turbidity. The stepwise approach remains two independent variables involved in multiple regression equation, color-spectrum and electrical conductivity with the value of   R-square equal to 0,97. Meaning that the two variables have the ability to explaining variations in turbidity up to 97 %.