numpy.ma.polyfit#

ma.polyfit(x, y, deg, rcond=None, full=False, w=None, cov=False)[source]#

Least squares polynomial fit to data with possible masked values.

This function is the equivalent of numpy.polyfit that takes masked values into account, see numpy.polyfit for details.

See also

numpy.polyfit: Equivalent function for ndarrays.

Notes

Any masked values in x are propagated to y, and vice-versa. A data point is excluded from the fit if either coordinate is masked.

Examples

>>> import numpy as np

Fit a line to data with a masked outlier in y:

>>> x = np.ma.array([0., 1., 2., 3., 4.])
>>> y = np.ma.array([1., 3., 5., 7., 999.], mask=[0, 0, 0, 0, 1])
>>> np.ma.polyfit(x, y, 1)
array([2., 1.])

Masking a value in x also excludes the corresponding y point:

>>> x = np.ma.array([0., 1., 999., 3., 4.], mask=[0, 0, 1, 0, 0])
>>> y = np.ma.array([1., 3., 5., 7., 9.])
>>> np.ma.polyfit(x, y, 1)
array([2., 1.])

Without masking, an outlier distorts the fit significantly:

>>> np.polyfit([0., 1., 2., 3., 4.], [1., 3., 5., 7., 999.], 1)
array([ 200., -197.])