An outgrowth of the relative frequency view of probability is the development of inductive statistical models. The statistical models used in the Earth sciences are many and varied. They range from simple techniques, such as the chi-square test, Student’s t-test, and the more versatile analysis of variance, all of which are used in testing the significance of differences between class data; through the more sophisticated techniques of regression and correlation, which are used for establishing relationships between variables; to complex multivariate models, such as multivariate regression, multivariate correlation, canonical correlation, multidimensional scaling, and principal component analysis, which enable relationships to be detected within sets of variables.
KeywordsPrincipal Component Analysis Drainage Basin Canonical Correlation Slope Gradient Principal Coordinate Analysis
Unable to display preview. Download preview PDF.