Showing variability in AGN by principal component analysis (PCA)

Principal component analysis (PCA) is a powerful tool for studying spectral variability. The technique consists of splitting relatively long exposures into a series of shorter-exposure spectra, and returns a minimal set of independent spectral shapes representing the variable components. If the initial spectra are made up of a linear sum of variable, uncorrelated and spectrally distinct physical components, the PCA will return detailed spectra of each variable component in a model independent way. This is a big advantage to analyze and study the origin of the observed variability without being limited by available spectral models (and by the systematic uncertainties that are inherent to any spectral analysis). We are applying the PCA analysis to several XMM-Newton observations from the brightest and most variable AGN with sufficiently long exposures. We shall present the most interesting results obtained so far.