Search results

  1. S

    PCA: PC scores or factor scores?

    Hi all, I noticed that some statistical softwares (SPSS, XLstat) label the axes of PCA biplots as "F1, F2,..." and call the score of each sample on each axis the "factor scores". Why isn't it PC1, PC2, ... and PC scores? I find it confusing (looks like results from a factor analysis). See...
  2. S

    Adding supplementary variables to clr-based PCA

    Hi all, I regularly work with compositional data, for which it is recommended to apply a clr-transformation before making a PCA. This is fine. Now, to explain the variance seen in the PCA, I need to project supplementay/explanatory variables that do not affect the shape of the PCA. Should the...
  3. S

    Linearly increasing trend in PLS residuals

    Hi everyone, I am currently using partial least square regressions to predict a variable from a set of 6 others. In general, residuals are randomly distributed, with sometimes a certain level of heteroscedasticity. For one dataset, however, the residuals are positively correlated to the variable...