The mathematical approach for PCA is to approximate
the data matrix** ***X*,**
**which has n objects and p variables, by two smaller matrices:
the scores matrix *T*
(n objects and d variables) and the loadings matrix *L
(d objects and p variables),
where *
(see Figure 1). One widely used algortihm for PCA is the NIPALS (**N**onlinear
**I**terative **P**artial **L**east **S**quares) algorithm.