RegionUncertainty

class sherpa.plot.RegionUncertainty

Bases: Confidence2D

The Region-Projection method.

Evaluate the statistic on a grid of points for two parameters, where the other thawed parameters are not changed from their current values.

Changed in version 4.16.1: Support for logarithmically-spaced grids has been improved.

Methods

calc(fit, par0, par1[, methoddict, cache])	Evaluate the statistic for the parameter range.
contour([overplot, clearwindow])
point(x, y[, overplot, clearwindow])	Draw a point at the given location.
prepare([min, max, nloop, delv, fac, log, ...])	Set the data to plot.

overcontour

See also

RegionProjection

Attributes Summary

conf_type	The type of confidence analysis.
contour_prefs	The preferences for the plot.
point_prefs	The preferences for the plot.

Methods Summary

calc(fit, par0, par1[, methoddict, cache])	Evaluate the statistic for the parameter range.
contour([overplot, clearwindow])
overcontour(*args, **kwargs)
point(x, y[, overplot, clearwindow])	Draw a point at the given location.
prepare([min, max, nloop, delv, fac, log, ...])	Set the data to plot.

Attributes Documentation

conf_type = 'uncertainty'

The type of confidence analysis.

contour_prefs: dict[str, Any] = {'alpha': None, 'colors': None, 'label': None, 'levels': None, 'linestyles': 'solid', 'linewidths': None, 'xlog': False, 'ylog': False}

The preferences for the plot.

point_prefs: dict[str, Any] = {'alpha': None, 'capsize': None, 'color': None, 'ecolor': None, 'label': None, 'linestyle': 'solid', 'linewidth': None, 'marker': '+', 'markerfacecolor': None, 'markersize': None, 'xerrorbars': False, 'xlog': False, 'yerrorbars': False, 'ylog': False}

The preferences for the plot.

Methods Documentation

calc(fit, par0, par1, methoddict=None, cache=True)

Evaluate the statistic for the parameter range.

This requires prepare to have been called, and must be called before contour is called.

Parameters:

fit

The Sherpa fit instance to use (defines the statistic and optimiser to use).

par0, par1

The parameters to iterate over.

See also

contour, prepare

Notes

This method is assumed to be over-ridden in derived classes, where it will perform the statistic calculations needed to create the visualization. This version should be called from these classes as it validates the fit and par arguments.

contour(overplot=False, clearwindow=True, **kwargs)

overcontour(*args, **kwargs)

point(x, y, overplot=True, clearwindow=False, **kwargs)

Draw a point at the given location.

Parameters:

x, y

The coordinates of the plot.

overplotbool, optional

If True then add the data to an existing plot, otherwise create a new plot.

clearwindowbool, optional

Should the existing plot area be cleared before creating this new plot (e.g. for multi-panel plots)?

**kwargs

These values are passed on to the plot backend, and must match the names of the keys of the object’s point_prefs dictionary.

prepare(min=None, max=None, nloop=(10, 10), delv=None, fac=4, log=(False, False), sigma=(1, 2, 3), levels=None, numcores=None)

Set the data to plot.

This defines the ranges over which the statistic will be calculated, but does not perform the evaluation.

Parameters:

min, maxsequence of number or None, optional

The minimum and maximum parameter values to used. If set then they must contain two elements, and if not then the range is calculated using the fac parameter.

nloopsequence of int, optional

The number of points at which to evaluate the statistic, where each value must be greater than 1. This is used when delv is set to None.

delvsequence of number or None, optional

The spacing of the parameter grids, and if set it must contain two values each greater than 0. This takes precedence over nloop.

facnumber, optional

Used when either min or max are not set. The parameter range in this case is taken to be fac times the separation of the covariance limits for the parameter (unless explicitly given).

logsequence of bool, optional

Should each parameter be evaluated on a logarithmically-spaced grid rather than a linearly-spaced one?

sigmasequence of number, optional

The sigma values at which to draw contours. This is only used if levels is set to None.

levelssequence of number or None, optional

The levels at which the contours are drawn. This over-rides the sigma setting.

numcoresint or None, optional

Should the parameter evaluation use multiple CPU cores if available?

See also

calc