CStatNegativePenalty

class sherpa.stats.CStatNegativePenalty(name: str = 'cstatnegativepenalty')[source] [edit on github]

Bases: CStat

CStat with penalty for negative model values.

The Cash statistic assumes that the model evaluation is always positive, as it requires evaluating the log of the model values. The CStat class replaces any model value that is 0 or negative by a small positive model (the trunc_value term).

The CStatNegativePenalty statistic instead applies a penalty to any model evaluation that leads to model values 0 or less, and this penalty increases as the values move further away from zero (in CStat the penalty is a constant term). The aim is to guide the optimizer back to models that are all positive.

The penalty term is (where M represents the model):

(sum_j M(j))

where j represents only those model values that are 0 or less.

When the model values are positive the statistic is the same as calculated by CStat.

Added in version 4.18.0.

Methods Summary

`calc_stat`(data, model)	Return the statistic value for the data and model.
`calc_staterror`(data)	Return the statistic error values for the data.
`goodness_of_fit`(statval, dof)	Return the reduced statistic and q value.

Methods Documentation

calc_stat(data: Data | DataSimulFit, model: Model) → tuple[float, ndarray] [edit on github]

Return the statistic value for the data and model.

Parameters:

data (sherpa.data.Data or sherpa.data.DataSimulFit) – The data set, or sets, to use.
model (sherpa.models.model.Model or sherpa.models.model.SimulFitModel) – The model expression, or expressions. If a sherpa.models.model.SimulFitModel is given then it must match the number of data sets in the data parameter.

Returns:

statval (number) – The value of the statistic.
fvec (array of numbers) – The per-bin “statistic” value.

static calc_staterror(data: ndarray) → ndarray [edit on github]

Return the statistic error values for the data.

Parameters:: data (scalar or 1D array of numbers) – The data values.
Returns:: staterror – The errors for the input data values (matches the data argument).
Return type:: scalar or array of numbers

goodness_of_fit(statval: float, dof: int) → tuple[Literal[None], Literal[None]] | tuple[float, float] [edit on github]

Return the reduced statistic and q value.

The reduced statisitc is conceptually simple, as it is just statistic / degrees-of-freedom, but it is not meaningful for all statistics, and it is only valid if there are any degrees of freedom.

Parameters:

statval (float) – The statistic value. It is assumed to be finite.
dof (int) – The number of degrees of freedom, which may be 0 or negative.

Returns:

rstat (float or NaN or None) – The reduced statistic. If the statistic does not support a goodness of fit then the return value is None. If it does then NaN is returned if either the number of degrees of freedom is 0 (or less), or the statistic value is less than 0.
qval (float or NaN or None) – The q value. If the statistic does not support a goodness of fit then the return values are None. If it does then NaN is returned if either the number of degrees of freedom is 0 (or less), or the statistic value is less than 0.