Beijing Normal Univ, Dept Astron, Beijing 100875, Peoples R China
; Tsinghua Univ, Ctr Astrophys, Beijing 100084, Peoples R China
; Chinese Acad Sci, Inst High Energy Phys, Key Lab Particle Astrophys, Beijing 100039, Peoples R China
Modern methods of spectral estimation based on parametric tin-le-series models are useful tools in power spectral analysis. We apply the autoregressive (AR) model to study quasi-periodic oscillations (QPOs) An empirical formula to estimate the expectation and standard deviation of the noise AR power densities is derived, which can be used to estimate the statistical significance of an apparent QPO peak in an AR spectrum. An iterative adding-noise algorithm in AR spectral analysis is proposed and applied to studying QPOs in the X-ray binary Cir X-1.