Alternative links to fulltext:
Abstract
Both principal component (PCA) and factor anal. (FA) reduce the dimensionality of multivariate biol. data of drugs by extg. uncorrelated "inner variables" related to the basis mechanisms of action. Owing to important model differences between PCA and FA, the choice of the method to be used depends on the type and the correlation structure of the data under consideration. Two examples, the 1st ...
Abstract
Both principal component (PCA) and factor anal. (FA) reduce the dimensionality of multivariate biol. data of drugs by extg. uncorrelated "inner variables" related to the basis mechanisms of action. Owing to important model differences between PCA and FA, the choice of the method to be used depends on the type and the correlation structure of the data under consideration. Two examples, the 1st typical of the PCA-, the 2nd of the FA-model, are presented with the intention to demonstrate the power and practical applicability of both approaches in multivariate QSAR work. The object components and factor values resulting from PCA and FA, resp., are correlated with mol. parameters to show the dependence of the "inner variables" on the structure of the drugs.
Altmetric