PCA for speed up ML models