Advanced Techniques Chemetrica also provides a range of routines for the more advanced user: Canonical Correlation Analysis (CCA) - the multivariate extension of univariate correlations.  Multiple linear regression, Linear Discriminant Analysis and MANOVA are special forms of CCA. Singular Value Decomposition (SVD).  SVD is used internally by many of Chemetrica's routines. It is also available to the user with a variety of scaling options, allowing access to other multivariate techniques using the SVD, such as Procrustes analysis. Multicollinearity Diagnostics.  Multivariate datasets usually show a high degree of correlation between the variables, which can lead to numerical instability in traditional regression.  These diagnostics let you see which variables are correlated, and how much damage is being done to an OLS regression as a result.  Variance Inflation Factors, Squared Multiple Correlations and Variance Decomposition Proportions are all calculated. Return to Chemetrica features page.