SIMBAD references

Applications of auto-regressive (AR) methods in asteroseismology are particularly interesting for high-resolution estimation of power spectra. We employ two kinds of AR processes, reflecting two different approaches to the problem: (1) a single AR process optimally fitted to the entire data set by the Maximum Entropy Method (MEM) principle; (2) several AR processes, each for a single gap, fitted by least squares to optimally represent data surrounding a given gap. In (1) we assume strictly stationary behaviour of the observed process and exploit its properties to the full extent, while in (2) assumptions are more flexible and enable us to treat with reasonable approximation signals with properties evolving over the interval of observation. We apply our methods to simulations and to observations of the variable stars HL Tau 76 and GD 154 (WET observations). The quality of different chunks of WET data being different, we account for these differences in quality by weighting the data.