Component analysis doesn't fundamentally reduce information, it just rotates it into a more informative basis. People usually drop vectors using the eigenvalues

Component analysis doesn't fundamentally reduce information, it just rotates it into a more informative basis. People usually drop vectors using the eigenvalues to do dimensionality reduction, but you don't have to do that.