The sparse principal component regression is computed. The regularization parameters are optimized by cross-validation.
