Mixed Data Fitter#

class MixedDataFitter(x_values: List | ndarray, y_values: List | ndarray, model_list: List[str], fitter_dictionary=None, max_iterations: int = 1000)[source]#

Bases: object

Class to fit a mixture of different models to data.

Parameters:

x_values – The x-values for the data.
y_values – The y-values for the data.
model_list – List of models to fit (e.g., LINE, GAUSSIAN, LOG_NORMAL)
max_iterations – The maximum number of iterations for fitting procedure.

Methods

`fit`(p0[, frozen])	Fit the data.
`get_fitted_curve`()	Gets the y-values from the fitted model.
`get_model_parameters`([model, errors])	Extracts parameters (and error) values for a specific model, or for all models if no model is specified.
`get_value_error_pair`([mean_values, std_values])	Retrieve the value/error pairs for the fitted parameters.
`plot_fit`([show_individuals, x_label, ...])	Plot the fitted models.

_create_model_function() → Callable[source]#

Creates a composite model function based on the specified models.

Returns:: A composite model for fitting.

_expected_param_count() → int[source]#

Calculates the expected number of parameters based on the model list.

Returns:: The number of parameters.

static _format_param(value, t_low=0.001, t_high=10000) → str[source]#

Formats the parameter value to scientific notation based on its magnitude.

Parameters:

value: float: The value of the parameter to be formatted.
t_low: float, optional: The lower bound below which the formatting should be applied to the value. Defaults to 0.001.
t_high: float, optional: The upper bound above which the formatting should be applied to the value. Defaults to 10,000.

Returns:

str:: A formatted string of the parameter value.

_get_bounds() → Tuple[ndarray, ndarray][source]#

Sets the bounds for each parameter based on the model list.

Returns:: Lower and upper bounds for the parameters.

_instantiate_fitter(model: str, return_values: str | List[str] = 'class')[source]#

Instantiate the fitter for the specified model and return requested values.

Parameters:

model – The model name as a string.
return_values – The specific attribute(s) or instance to return. Options are ‘class’, ‘n_par’, and ‘bounds’. Can be a string (for one value) or a list of strings.

Returns:

The requested values as a single value or a tuple.

Raises:

ValueError – If the model is not recognized or return_values are invalid.

_parameter_extractor(values: ndarray) → dict[source]#

Extracts the parameters for each model in the model list.

Parameters:: values – The values from which the model dictionary is to be extracted.
Returns:: A dictionary where the keys are model names and the values are lists of parameters/error values.

_params() → ndarray[source]#

Store the fitted parameters of the fitted model.

Returns:

np.ndarray: The parameters obtained after performing the fit.

Raises:

RuntimeError: If the fit has not been performed yet (i.e., self.params is None).

Notes

This method assumes that the fitting process assigns values to self.params.

_plot_individual_fitter(plotter)[source]#

Plot the individual fitters function.

Parameters:: plotter – The plotting axis object

_standard_errors() → ndarray[source]#

Store the standard errors of the fitted parameters.

Returns:

np.ndarray: An array containing the standard errors of the fitted parameters.

Raises:

RuntimeError: If the fit has not been performed yet (i.e., self.covariance is None).

fit(p0: List | ndarray, frozen: int | List[int] | None = None)[source]#

Fit the data.

Parameters:

p0 (Union[List, np.ndarray]) – Initial guess for the fitted parameters.
frozen (Union[int, List[int]]) – Parameter number of list of parameter numbers to freeze the value of.

Raises:

ValueError – If the length of the initial guess is not equal to the expected parameter count.

get_fitted_curve() → ndarray[source]#

Gets the y-values from the fitted model.

Returns:: The y-values from the fitted model
Raises:: ValueError – If the model has not been fitted yet.

get_model_parameters(model: str | None = None, errors: bool = False)[source]#

Extracts parameters (and error) values for a specific model, or for all models if no model is specified.

Parameters:

model – Model name to extract parameters for. If unspecified, extracts parameters for all models. Defaults to None.
errors – If True, includes the errors in the returned output. Defaults to False.

Returns:

A dictionary containing:

”parameters”: Nested dictionary of parameter values for each model if get_errors is True.
”errors”: Nested dictionary of errors for each model (if get_errors=True).

Otherwise, returns just the parameters directly.

get_value_error_pair(mean_values: bool = True, std_values: bool = False) → ndarray[source]#

Retrieve the value/error pairs for the fitted parameters.

This method provides the fitted parameter values and their corresponding standard errors as a combined array or individually based on the input flags.

Parameters:

mean_valuesbool, optional: If True, return only the values of the fitted parameters. Defaults to True.
std_valuesbool, optional: If True, return only the standard errors of the fitted parameters. Defaults to False.

Returns:

np.ndarray

If mean_values and std_values are both True: A 2D array of shape (n_parameters, 2),
where each row is [value, error].
If mean_values is True and std_values is False: A 1D array of parameter values.
If std_values is True and mean_values is False: A 1D array of standard errors.
If both flags are False: An error message.

Raises:

ValueError: If both mean_values and std_values are False.

Plot the fitted models.

Parameters:

show_individuals: bool, optional: Whether to show individually fitted models or not.
x_label: str, optional: The label for the x-axis.
y_label: str, optional: The label for the y-axis.
title: str, optional: The title for the plot.
data_label: str, optional: The label for the data.
axis: Axes, optional: Axes to plot instead of the entire figure. Defaults to None.

Returns:

plotter: The plotter handle for the drawn plot.

Currently the MixedDataFitter supports fitting the following models together,

Recommended Import#

from pymultifit.fitters import MixedDataFitter

Full Import#

from pymultifit.fitters.mixed_f import MixedDataFitter