It accomplishes this reduction by identifying directions, called principal components, along which the variation in the data is maximal. By using ... samples, PCA can reduce the dimensionality ...