All sparse PCA models are wrong, but some are useful. Part I: Computation of scores, residuals and explained variance | Publicación