astra.contrib.thepayne.test

Module Contents

Functions

_predict_stellar_spectrum(unscaled_labels, weights, biases)
_redshift(dispersion, flux, radial_velocity)
load_state(path)
test(wavelength, flux, ivar, neural_network_coefficients, scales, model_wavelength, label_names, initial_labels=None, radial_velocity_tolerance=None, **kwargs) Use a pre-trained neural network to estimate the stellar labels for the given spectrum.
get_chi_sq(expectation, y, y_err, L)
astra.contrib.thepayne.test.LARGE = 1000.0
astra.contrib.thepayne.test.c
astra.contrib.thepayne.test.sigmoid
astra.contrib.thepayne.test._predict_stellar_spectrum(unscaled_labels, weights, biases)
astra.contrib.thepayne.test._redshift(dispersion, flux, radial_velocity)
astra.contrib.thepayne.test.load_state(path)
astra.contrib.thepayne.test.test(wavelength, flux, ivar, neural_network_coefficients, scales, model_wavelength, label_names, initial_labels=None, radial_velocity_tolerance=None, **kwargs)

Use a pre-trained neural network to estimate the stellar labels for the given spectrum.

Parameters:
  • wavelength – The wavelength array of the observed spectrum.
  • flux – The observed (or pseudo-continuum-normalised) fluxes.
  • ivar – The inverse variances of the fluxes.
  • neural_network_coefficients – A two-length tuple containing the weights of the neural network, and the biases.
  • scales – The lower and upper scaling value used for the labels.
  • initial_labels – [optional] The initial labels to optimize from. By default this will be set at the center of the training set labels.
  • radial_velocity_tolerance – [optional] Supply a radial velocity tolerance to fit simulatenously with stellar parameters. If None is given then no radial velocity will be fit. If a float/integer is given then any radial velocity +/- that value will be considered. Alternatively, a (lower, upper) bound can be given.
Returns:

A three-length tuple containing the optimized parameters, the covariance matrix, and a metadata dictionary.
astra.contrib.thepayne.test.get_chi_sq(expectation, y, y_err, L)