Pincipal Component Analysis
Hi Julytly,
Principal component analysis does not play much of a role in my work. When optimizing a sampling protocol or evaluating a bulk sampling program, the most test results (and degrees of freedom) are obtained at the analytical stage and the fewest degrees of freedom at the primary sample selection stage. In other words, test results can be ordered in a hierarchy of three classes, namely, sample selection, sample preparation and analysis. Calculations are simple and matrix transformations are not required. Wish I could assist in identifying the largest variance in your matrix.
Kind regards,
Jan W Merks ■
PCA question
Hi,
I am now doing PCA for a sample matrix. I used matlab command princomp, which returns PC(eigenvector), SCORE(zscore), LATENT(eigenvalue). When I discussed with my colleages, my feeling is that there is more than one way to do PCA. I am confused whether the first and second columns of score can represent the first 2 principal components. Could anyone give me some suggestions on this?
Many thanks. ■