Multivariate statistical techniques and water quality assessment: Discourse and review on some analytical models

Regular monitoring and comprehensive assessment of water quality and its associated processes require sophisticated analytical models to reveal concealed instruments controlling their properties. This information is essential to design monitoring frameworks and sustainable management of the water resources. Intelligent data analysis techniques like multivariate statistical models can greatly assist in water quality management programs. This paper provides basic knowledge of the five multivariate data mining approaches, namely, cluster analysis, principal component analysis, factor analysis, multiple linear regression analysis and discriminant analysis, and highlights their applications in the characterization and classification of the surface water quality. The applicability of multivariate tools for the river basin management is the principal focus of this communication. Furthermore, this literature review also presents some of the basic concepts of the newly employed source apportionment receptor modeling technique involving multiple linear regression (MLR) and absolute principal component scores (APCS-MLR model) for extensive water quality assessment.