Multivariate Statistical Anlaysis in Water Quality

Multivariate statistical techniques are the application of statistics to simultaneous observations and can include the analysis of more than one outcome (dependent) variable. Good multivariate analysis starts with exploratory and graphical analyses to reveal potential relations in the data and to highlight potential outliers. First, this presentation will discuss how to extend univariate and bivariate methods for graphical analysis to multivariate data, as well as methods unique to multivariate data. Second, multivariate outlier detection will be presented. Third, there will be a brief discussion of multivariate statistical analysis methods, such as multiple regression, principal component analysis, and cluster analysis, including examples and suggestions as to when one might want to use these techniques.