property Data.relative_error#

Relative error of the data.

This can be set using an array of the same size as the data (e.g. if you want to assign a different relative error to each datum) or as a scalar if you would like to assign a the same relative error to all data.

The standard_deviation is constructed as follows:

np.sqrt( (relative_error * np.abs(dobs))**2 + noise_floor**2 )

For example, if you set

>>> data = Data(survey, dobs=dobs)
>>> data.relative_error = 0.05

then the contribution to the standard_deviation is equal to

>>> data.relative_error * np.abs(data.dobs)
None or float or numpy.ndarray