Assessment of drinking water quality using principal component analysis and partial least square discriminant analysis: a case study at water treatment plants, Selangor

Abstract

This study characterizes the drinking water quality on 28 water treatmentt plants in Selangor from 2009 to 2012 using multivariate techniques. The objectives of this study are to analyze the quality of collected drinking water and to detect the source of pollution for the most revealing  parameters.  The  Partial  Least  Square  Discriminant  Analysis (PLS-DA)  model showed a high correlation matrix of analysis for physicochemical quality of two types of water with  99.43% significant  value.  The classification  matrix  accuracy of the principal component  analysis  (PCA) highlighted  13  significant  physico-chemical water quality parameters and 14 significant heavy metal parameters. PCA was carried out to identify the origin and source of pollution of each water quality parameters. Therefore, this study proves that chemometric method is the principle way to characterize the drinking water quality.

Keywords: partial least square, discriminant analysis; principal component analysis; drinking water quality