get_chisq

mufasa.UltraCube.get_chisq(cube, model, expand=20, reduced=True, usemask=True, mask=None)

Calculate the chi-squared or reduced chi-squared value for a spectral cube.

This function computes the chi-squared goodness-of-fit statistic by comparing the observed data in the cube with a provided model. Optionally, the calculation can be restricted to masked regions and expanded spectral regions.

Parameters:

• cube (SpectralCube) – The observed spectral cube containing the data to compare against the model.

• model (numpy.ndarray) – A 3D array representing the model cube, which must have the same shape as the input cube.

• expand (int, optional) – Number of spectral channels to expand the mask region. Default is 20.

• reduced (bool, optional) – If True, compute the reduced chi-squared value by normalizing with the degrees of freedom. If False, compute the standard chi-squared value. Default is True.

• usemask (bool, optional) – If True, apply a mask to exclude invalid regions from the calculation. If no mask is provided, regions where the model is zero are excluded. Default is True.

• mask (numpy.ndarray, optional) – A 3D boolean array specifying regions to include in the calculation. If None, a mask is automatically derived from the model.

Returns:

A 2D array representing the chi-squared (or reduced chi-squared) value at each spatial pixel in the cube.

Return type:

numpy.ndarray

Notes

• The expand parameter allows users to extend the region of interest by a specified number of spectral channels around the model.

• The reduced chi-squared value is normalized by the number of degrees of freedom.

• The residuals between the cube and model are weighted using the mask.

Raises:

ValueError – If the cube and model dimensions do not match.

Examples

>>> from spectral_cube import SpectralCube
>>> import numpy as np
>>> cube = SpectralCube.read("example_cube.fits")
>>> model = np.random.random(cube.shape)
>>> chisq = get_chisq(cube, model, expand=10, reduced=True)
>>> print(chisq)