I use very often Win ISI, and I like to compare how the algorithms work in Win ISI compares with other software and especially with the R packages. Win ISI use SVD as the method to calculate the scores and loadings matrix. Resemble use also this algorithm.In the post "Tutorials with Resemble (part3)", we saw the scores map for the first and second PC.
In the post “Importing NIRsoil spectra fromResemble into WinISI”, we saw how to import and work with NIRsoil spectra in Win ISI.
Win ISI has the function “Create a Score file from a Spectra file”, where we calculate the PCs and also detect the outliers, where we can plot the sample scores maps.
And we can also over plot the validation samples with the training samples in order to check if the validation samples fit in the space defined by the training samples.
I show here the plot from Resemble with the training samples in blue color and the validation samples in red.
Next picture show the sample plot but in Win ISI with the training samples in blue and the validation samples in green. As you can see the are the same.