ilex.pyfit

class PyfitLikelihood(*args: Any, **kwargs: Any)[source]

calc_sigma()[source]

get_sigma()[source]

log_likelihood()[source]: Evaluate the log liklihood of function

property model_parameters

propogated_sigma()[source]

property sigma

class fit(**kwargs)[source]

General fitting class, model a function using either bayesian inference using BILBY, or least squares using scipy.curve_fit(). This class has a handful of useful quality of life features, including model plotting and statistics such as chi2.

x

X data

Type:: np.ndarray

y

Y data

Type:: np.ndarray

xerr

X data error

Type:: np.ndarray

yerr

Y data error

Type:: np.ndarray

func

Callable python function to model/fit to

Type:: __func__

method

Method to fit

[bayesian] - Use bayesian inference (see https://lscsoft.docs.ligo.org/bilby/)

[least squares] - use least squares method (see https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.curve_fit.html)

Type:: str

fit_keywords

Dictionary of keywords to apply to fitting method

Type:: Dict

prior

Priors for modelling

Type:: Dict

static

Priors to keep constant during modelling

Type:: Dict

posterior

Sampled posteriors

Type:: Dict

likelihood

Likelihood class used for bayesian inference

Type:: Bilby.Likelihood

keys

parameter names to be sampled

Type:: List

chi2

Chi squared statistic

Type:: float

chi2err

Chi squared error

Type:: float

rchi2

reduced Chi squared

Type:: float

rchi2err

reduced Chi squared error

Type:: float

nfitpar

number of fitted parameters

Type:: int

dof

degrees of freedom

Type:: int

p

p-value statistic

Type:: float

bic

bayesian information criterion

Type:: float

bic_err

bic error

Type:: float

residuals

if true, plot residuals when .plot() is called

Type:: bool

plotPosterior

if true, save image of posterior corner plot

Type:: bool

fit(redo=False, **kwargs)[source]

Fit to function

Parameters:: redo (bool) – redo fitting (in case BILBY is being used, remove cached results of any previous fit), default is False

get_mean_err(func=True)[source]

Get mean errors of posterior

Parameters:: func (bool, optional) – if true, retrieve errors of posteriors defined in function
Returns:: _errs – dictionary of errors for posteriors
Return type:: dict

get_model(x=None)[source]

Get model fit

Parameters:

x (array-like) – datapoints to evaluate model, if None will use x values already given to current instance

Returns:

x (array-like) – x data
y (array-like) – y data - evaluated model values

get_post_val(func=True)[source]: Get values/betfit of posterior

get_posteriors(func=True)[source]

Get posteriors

Parameters:: func (bool) – if true, only retrieve posteriors of parameters of function to be modelled, default is True
Returns:: posteriors – Dictionary of posterior median/mean best values
Return type:: dict

get_priors(func=True)[source]

get priors and statics

Parameters:: func (bool) – if true, only retrieve priors of parameters of function to be modelled, default is True

plot(show=True, filename=None, **ax_kw)[source]

Plot fitted model and data

Parameters:

show (bool, optional) – if true, display plot
filename (str) – name of output image file, default is None
**ax_kw (Dict) – keyword parameters for plotting

Returns:

fig – figure instance

Return type:

plt._figure_

set(**kwargs)[source]

set_posterior(name, val, plus, minus)[source]

Set posterior

Parameters:

name (str) – name of posterior
val (float) – bestfit value/median
plus (float) – positive std/err
minus (float) – negative std/err

set_prior(name, pr)[source]

Set prior

Parameters:

name (str) – name of parameter
pr (float or 2-element list) – prior value

property sigma

stats()[source]: Print statistics