asterion.models#

Probabilistic models for asteroseismic oscillation mode frequencies.

Module Contents#

class GlitchModel(nu_max, delta_nu, teff=None, epsilon=None, seed=0, window_width='full')[source]#

Bases: Model

Asteroseismic glitch model.

\[\nu_\mathrm{obs} \sim \mathcal{GP}(m(n), k(n, n') + \sigma^2\mathcal{I})\]

Where the mean function is,

\[\begin{split}m(n) &= \nu_\mathrm{bkg} + \delta_\mathrm{He} + \delta_\mathrm{CZ},\\ \nu_\mathrm{bkg} &= f_\mathrm{bkg}(n),\\ \delta_\mathrm{He} &= f_\mathrm{He}(\nu_\mathrm{bkg}),\\ \delta_\mathrm{CZ} &= f_\mathrm{CZ}(\nu_\mathrm{bkg}),\end{split}\]

and the kernel function is,

\[k(n, n') = \sigma_k^2 \exp\left( - \frac{(n' - n)^2}{l^2} \right).\]
Parameters
  • n (array_like) – Radial order of model observations.

  • nu_max (dist_like) – Prior on the frequency at maximum power.

  • delta_nu (dist_like) – Prior on the large frequency separation.

  • teff (dist_like, optional) – Prior on the effective temperature. This is used for estimating a prior on the glitch acoustic depths. If None (default), a prior of Normal(5000, 700) is assumed.

  • epsilon (dist_like, optional) – Prior on the asymptotic phase parameter.

  • num_pred (int) – The number of points in radial order for which to make predictions.

  • seed (int) – The seed used to generate samples from the prior on the glitch periods (acoustic depths) tau_he and tau_cz.

  • window_width (float) – The number of delta_nu either side of nu_max over which to average the helium glitch amplitude for the parameter ‘he_amplitude’.

n#

Radial order of model observations.

Type

numpy.ndarray

n_pred#

Radial order of model predictions.

Type

numpy.ndarray

background#

Prior on the background function.

Type

Prior

he_glitch#

Prior on the helium glitch function.

Type

Prior

cz_glitch#

Prior on the base of convective zone glitch function.

Type

Prior

window_width#

The number of delta_nu either side of nu_max over which to average the glitch amplitudes. If string, ‘full’, the window is chosen over the entire range in frequency.

Type

str or float

class Model(*args, **kwargs)[source]#

Bases: asterion.priors.Prior

Model class.

A model is a probabilistic object which may be given to Inference. It does not need to return anything during inference, but should have at least one observed sample sites.