Author(s): Md Mahmudul Haque; Ataur Rahman; Dharma Hagare; Golam Kibria
Linked Author(s):
Keywords: Principal component analysis; Multiple regression; Water demand; Predictor variables; Multicollinearity
Abstract: In regression based water demand forecast modelling, identification of suitable predictor variables is an important step. A variety of factors associated with climate change, population growth, socio-economic conditions and policy issues generally influence urban water demand. These variables are often correlated with each other, which may create problem in building appropriate water demand forecast model. Therefore, selection of the appropriate predictor variables is important for accurate prediction of future water demand which assists in planning and designing efficient water utility system. In this study, principal component analysis (PCA) was carried out using 10 independent variables in relation to per dwelling water consumption level in the Blue Mountains Water Supply systems in NSW, Australia. Here, principal component scores and multiple linear regression analysis were combined together to develop a principal component regression (PCR) model to predict the future water demand. PCA is a multivariate statistical analysis to reduce a large number of correlated variables into a smaller number of independent components to reduce the dimension of the problem. The results showed that the significant predictor variables in water demand forecasting in the Blue Mountain regions were monthly total rainfall, number of rain days in month, monthly mean maximum temperature, monthly total evaporation, monthly mean daily global solar exposure, water price and water conservation savings. It was also found that the developed PCR model were able to predict future water demand with a high degree of accuracy with an average relative error value of 3. 47% . Moreover, it was found that PCR model could be used to eliminate the multicollinearity problem and to reduce the dimension of the variables in the multiple linear regression model.
Year: 2013