Model Selection for Cepheid Star Oscillations

摘要: SUMMARY Cepheid variables are a class of pulsating variable stars with the useful property that their periods variability strongly correlated absolute luminosity. Once this relationship has been calibrated, knowledge period gives This makes these as “standard candles” for estimating distances in universe. paper updates and expands work reported by Jefferys Barnes (1999). We consider fully Bayesian inference using reversible-jump MCMC simulation takes data photometric velocity information output posterior physical such luminosity star, its distance, radius, other parameters. model photometry velocities Fourier polynomials an unknown or selectable number terms; connected nonlinear relations involving parameters interest. From amongst models varying numbers coefficients we select highest probability, obtain on averaged over models, weights proportional to probabilities models. discuss issues concerning priors, effectiveness sampling, practical results our research program. briefly alternative wavelets instead polynomials, approaches priors.