https://github.com/lmfit/lmfit-py
Tip revision: 7fec373139313d7cc6c2b56cc3ea98e54decb82c authored by Matt Newville on 26 November 2018, 15:25:27 UTC
Merge pull request #519 from reneeotten/PR512
Merge pull request #519 from reneeotten/PR512
Tip revision: 7fec373
minimizer.py
"""Simple minimizer is a wrapper around scipy.leastsq, allowing a user to build
a fitting model as a function of general purpose Fit Parameters that can be
fixed or varied, bounded, and written as a simple expression of other Fit
Parameters.
The user sets up a model in terms of instance of Parameters and writes a
function-to-be-minimized (residual function) in terms of these Parameters.
Original copyright:
Copyright (c) 2011 Matthew Newville, The University of Chicago
See LICENSE for more complete authorship information and license.
"""
from collections import namedtuple
from copy import deepcopy
import multiprocessing
import numbers
import warnings
import numpy as np
from numpy import ndarray, ones_like, sqrt
from numpy.dual import inv
from numpy.linalg import LinAlgError
from scipy.optimize import basinhopping as scipy_basinhopping
from scipy.optimize import brute as scipy_brute
from scipy.optimize import differential_evolution, least_squares
from scipy.optimize import leastsq as scipy_leastsq
from scipy.optimize import minimize as scipy_minimize
from scipy.stats import cauchy as cauchy_dist
from scipy.stats import norm as norm_dist
from scipy.version import version as scipy_version
import six
import uncertainties
from ._ampgo import ampgo
from .parameter import Parameter, Parameters
# check for EMCEE
try:
import emcee
HAS_EMCEE = True
EMCEE_VERSION = int(emcee.__version__[0])
except ImportError:
HAS_EMCEE = False
# check for pandas
try:
import pandas as pd
from pandas import isnull
HAS_PANDAS = True
except ImportError:
HAS_PANDAS = False
isnull = np.isnan
# check for numdifftools
try:
import numdifftools as ndt
HAS_NUMDIFFTOOLS = True
except ImportError:
HAS_NUMDIFFTOOLS = False
# define the namedtuple here so pickle will work with the MinimizerResult
Candidate = namedtuple('Candidate', ['params', 'score'])
def asteval_with_uncertainties(*vals, **kwargs):
"""Calculate object value, given values for variables.
This is used by the uncertainties package to calculate the
uncertainty in an object even with a complicated expression.
"""
_obj = kwargs.get('_obj', None)
_pars = kwargs.get('_pars', None)
_names = kwargs.get('_names', None)
_asteval = _pars._asteval
if (_obj is None or _pars is None or _names is None or
_asteval is None or _obj._expr_ast is None):
return 0
for val, name in zip(vals, _names):
_asteval.symtable[name] = val
return _asteval.eval(_obj._expr_ast)
wrap_ueval = uncertainties.wrap(asteval_with_uncertainties)
def eval_stderr(obj, uvars, _names, _pars):
"""Evaluate uncertainty and set .stderr for a parameter `obj`.
Given the uncertain values `uvars` (a list of uncertainties.ufloats), a
list of parameter names that matches uvars, and a dict of param objects,
keyed by name.
This uses the uncertainties package wrapped function to evaluate the
uncertainty for an arbitrary expression (in obj._expr_ast) of parameters.
"""
if not isinstance(obj, Parameter) or getattr(obj, '_expr_ast', None) is None:
return
uval = wrap_ueval(*uvars, _obj=obj, _names=_names, _pars=_pars)
try:
obj.stderr = uval.std_dev
# TODO: do not use bare except
except:
obj.stderr = 0
class MinimizerException(Exception):
"""General Purpose Exception."""
def __init__(self, msg):
Exception.__init__(self)
self.msg = msg
def __str__(self):
return "{}".format(self.msg)
class AbortFitException(MinimizerException):
"""Raised when a fit is aborted by the user."""
pass
SCALAR_METHODS = {'nelder': 'Nelder-Mead',
'powell': 'Powell',
'cg': 'CG',
'bfgs': 'BFGS',
'newton': 'Newton-CG',
'lbfgsb': 'L-BFGS-B',
'l-bfgsb': 'L-BFGS-B',
'tnc': 'TNC',
'cobyla': 'COBYLA',
'slsqp': 'SLSQP',
'dogleg': 'dogleg',
'trust-ncg': 'trust-ncg',
'differential_evolution': 'differential_evolution'}
# FIXME: update this when incresing the minimum scipy version
major, minor, micro = scipy_version.split('.', 2)
if (int(major) >= 1 and int(minor) >= 1):
SCALAR_METHODS.update({'trust-constr': 'trust-constr'})
if int(major) >= 1:
SCALAR_METHODS.update({'trust-exact': 'trust-exact',
'trust-krylov': 'trust-krylov'})
def reduce_chisquare(r):
"""Reduce residual array to scalar (chi-square).
Calculate the chi-square value from the residual array `r`: (r*r).sum()
Parameters
----------
r : numpy.ndarray
Residual array.
Returns
-------
float
Chi-square calculated from the residual array
"""
return (r*r).sum()
def reduce_negentropy(r):
"""Reduce residual array to scalar (negentropy).
Reduce residual array `r` to scalar using negative entropy and the normal
(Gaussian) probability distribution of `r` as pdf:
(norm.pdf(r)*norm.logpdf(r)).sum()
since pdf(r) = exp(-r*r/2)/sqrt(2*pi), this is
((r*r/2 - log(sqrt(2*pi))) * exp(-r*r/2)).sum()
Parameters
----------
r : numpy.ndarray
Residual array.
Returns
-------
float
Negative entropy value calculated from the residual array
"""
return (norm_dist.pdf(r)*norm_dist.logpdf(r)).sum()
def reduce_cauchylogpdf(r):
"""Reduce residual array to scalar (cauchylogpdf).
Reduce residual array `r` to scalar using negative log-likelihood and a
Cauchy (Lorentzian) distribution of `r`:
-scipy.stats.cauchy.logpdf(r)
(where the Cauchy pdf = 1/(pi*(1+r*r))). This gives greater
suppression of outliers compared to normal sum-of-squares.
Parameters
----------
r : numpy.ndarray
Residual array.
Returns
-------
float
Negative entropy value calculated from the residual array
"""
return -cauchy_dist.logpdf(r).sum()
class MinimizerResult(object):
r"""The results of a minimization.
Minimization results include data such as status and error messages,
fit statistics, and the updated (i.e., best-fit) parameters themselves
in the :attr:`params` attribute.
The list of (possible) `MinimizerResult` attributes is given below:
Attributes
----------
params : :class:`~lmfit.parameter.Parameters`
The best-fit parameters resulting from the fit.
status : int
Termination status of the optimizer. Its value depends on the
underlying solver. Refer to `message` for details.
var_names : list
Ordered list of variable parameter names used in optimization, and
useful for understanding the values in :attr:`init_vals` and
:attr:`covar`.
covar : numpy.ndarray
Covariance matrix from minimization, with rows and columns
corresponding to :attr:`var_names`.
init_vals : list
List of initial values for variable parameters using :attr:`var_names`.
init_values : dict
Dictionary of initial values for variable parameters.
nfev : int
Number of function evaluations.
success : bool
True if the fit succeeded, otherwise False.
errorbars : bool
True if uncertainties were estimated, otherwise False.
message : str
Message about fit success.
ier : int
Integer error value from :scipydoc:`optimize.leastsq` (`leastsq` only).
lmdif_message : str
Message from :scipydoc:`optimize.leastsq` (`leastsq` only).
nvarys : int
Number of variables in fit: :math:`N_{\rm varys}`.
ndata : int
Number of data points: :math:`N`.
nfree : int
Degrees of freedom in fit: :math:`N - N_{\rm varys}`.
residual : numpy.ndarray
Residual array :math:`{\rm Resid_i}`. Return value of the objective
function when using the best-fit values of the parameters.
chisqr : float
Chi-square: :math:`\chi^2 = \sum_i^N [{\rm Resid}_i]^2`.
redchi : float
Reduced chi-square:
:math:`\chi^2_{\nu}= {\chi^2} / {(N - N_{\rm varys})}`.
aic : float
Akaike Information Criterion statistic:
:math:`N \ln(\chi^2/N) + 2 N_{\rm varys}`.
bic : float
Bayesian Information Criterion statistic:
:math:`N \ln(\chi^2/N) + \ln(N) N_{\rm varys}`.
flatchain : pandas.DataFrame
A flatchain view of the sampling chain from the `emcee` method.
Methods
-------
show_candidates
Pretty_print() representation of candidates from the `brute` method.
"""
def __init__(self, **kws):
for key, val in kws.items():
setattr(self, key, val)
@property
def flatchain(self):
"""Show flatchain view of the sampling chain from `emcee` method."""
if hasattr(self, 'chain'):
if HAS_PANDAS:
if len(self.chain.shape) == 4:
return pd.DataFrame(self.chain[0, ...].reshape((-1, self.nvarys)),
columns=self.var_names)
elif len(self.chain.shape) == 3:
return pd.DataFrame(self.chain.reshape((-1, self.nvarys)),
columns=self.var_names)
else:
raise NotImplementedError('Please install Pandas to see the '
'flattened chain')
else:
return None
def show_candidates(self, candidate_nmb='all'):
"""Show pretty_print() representation of candidates from `brute` method.
Showing candidates (default is 'all') or the specified candidate-#
from the `brute` method.
Parameters
----------
candidate_nmb : int or 'all'
The candidate-number to show using the :meth:`pretty_print` method.
"""
if hasattr(self, 'candidates'):
try:
candidate = self.candidates[candidate_nmb]
print("\nCandidate #{}, chisqr = "
"{:.3f}".format(candidate_nmb, candidate.score))
candidate.params.pretty_print()
except IndexError:
for i, candidate in enumerate(self.candidates):
print("\nCandidate #{}, chisqr = "
"{:.3f}".format(i, candidate.score))
candidate.params.pretty_print()
def _calculate_statistics(self):
"""Calculate the fitting statistics."""
self.nvarys = len(self.init_vals)
if isinstance(self.residual, ndarray):
self.chisqr = (self.residual**2).sum()
self.ndata = len(self.residual)
self.nfree = self.ndata - self.nvarys
else:
self.chisqr = self.residual
self.ndata = 1
self.nfree = 1
self.redchi = self.chisqr / max(1, self.nfree)
# this is -2*loglikelihood
_neg2_log_likel = self.ndata * np.log(self.chisqr / self.ndata)
self.aic = _neg2_log_likel + 2 * self.nvarys
self.bic = _neg2_log_likel + np.log(self.ndata) * self.nvarys
class Minimizer(object):
"""A general minimizer for curve fitting and optimization."""
_err_nonparam = ("params must be a minimizer.Parameters() instance or list "
"of Parameters()")
_err_maxfev = ("Too many function calls (max set to %i)! Use:"
" minimize(func, params, ..., maxfev=NNN)"
"or set leastsq_kws['maxfev'] to increase this maximum.")
def __init__(self, userfcn, params, fcn_args=None, fcn_kws=None,
iter_cb=None, scale_covar=True, nan_policy='raise',
reduce_fcn=None, calc_covar=True, **kws):
"""
Parameters
----------
userfcn : callable
Objective function that returns the residual (difference between
model and data) to be minimized in a least-squares sense. This
function must have the signature::
userfcn(params, *fcn_args, **fcn_kws)
params : :class:`~lmfit.parameter.Parameters`
Contains the Parameters for the model.
fcn_args : tuple, optional
Positional arguments to pass to `userfcn`.
fcn_kws : dict, optional
Keyword arguments to pass to `userfcn`.
iter_cb : callable, optional
Function to be called at each fit iteration. This function should
have the signature::
iter_cb(params, iter, resid, *fcn_args, **fcn_kws)
where `params` will have the current parameter values, `iter`
the iteration number, `resid` the current residual array, and `*fcn_args`
and `**fcn_kws` are passed to the objective function.
scale_covar : bool, optional
Whether to automatically scale the covariance matrix (default is True).
nan_policy : str, optional
Specifies action if `userfcn` (or a Jacobian) returns NaN
values. One of:
- 'raise' : a `ValueError` is raised
- 'propagate' : the values returned from `userfcn` are un-altered
- 'omit' : non-finite values are filtered
reduce_fcn : str or callable, optional
Function to convert a residual array to a scalar value for the scalar
minimizers. Optional values are (where `r` is the residual array):
- None : sum of squares of residual [default]
= (r*r).sum()
- 'negentropy' : neg entropy, using normal distribution
= rho*log(rho).sum()`, where rho = exp(-r*r/2)/(sqrt(2*pi))
- 'neglogcauchy': neg log likelihood, using Cauchy distribution
= -log(1/(pi*(1+r*r))).sum()
- callable : must take one argument (`r`) and return a float.
calc_covar : bool, optional
Whether to calculate the covariance matrix (default is True) for
solvers other than `leastsq` and `least_squares`. Requires the
`numdifftools` package to be installed.
**kws : dict, optional
Options to pass to the minimizer being used.
Notes
-----
The objective function should return the value to be minimized. For
the Levenberg-Marquardt algorithm from :meth:`leastsq` or
:meth:`least_squares`, this returned value must be an array, with a
length greater than or equal to the number of fitting variables in
the model. For the other methods, the return value can either be a
scalar or an array. If an array is returned, the sum of squares of
the array will be sent to the underlying fitting method, effectively
doing a least-squares optimization of the return values. If the
objective function returns non-finite values then a `ValueError`
will be raised because the underlying solvers cannot deal with them.
A common use for the `fcn_args` and `fcn_kws` would be to pass in
other data needed to calculate the residual, including such things
as the data array, dependent variable, uncertainties in the data,
and other data structures for the model calculation.
"""
self.userfcn = userfcn
self.userargs = fcn_args
if self.userargs is None:
self.userargs = []
self.userkws = fcn_kws
if self.userkws is None:
self.userkws = {}
self.kws = kws
self.iter_cb = iter_cb
self.calc_covar = calc_covar
self.scale_covar = scale_covar
self.nfev = 0
self.nfree = 0
self.ndata = 0
self.ier = 0
self._abort = False
self.success = True
self.errorbars = False
self.message = None
self.lmdif_message = None
self.chisqr = None
self.redchi = None
self.covar = None
self.residual = None
self.reduce_fcn = reduce_fcn
self.params = params
self.jacfcn = None
self.nan_policy = nan_policy
@property
def values(self):
"""Return Parameter values in a simple dictionary."""
return {name: p.value for name, p in self.result.params.items()}
def __residual(self, fvars, apply_bounds_transformation=True):
"""Residual function used for least-squares fit.
With the new, candidate values of `fvars` (the fitting variables),
this evaluates all parameters, including setting bounds and
evaluating constraints, and then passes those to the user-supplied
function to calculate the residual.
Parameters
----------
fvars : numpy.ndarray
Array of new parameter values suggested by the minimizer.
apply_bounds_transformation : bool, optional
Whether to apply lmfits parameter transformation to constrain
parameters (default is True). This is needed for solvers without
inbuilt support for bounds.
Returns
-------
residual : numpy.ndarray
The evaluated function values for given `fvars`.
"""
params = self.result.params
if fvars.shape == ():
fvars = fvars.reshape((1,))
if apply_bounds_transformation:
for name, val in zip(self.result.var_names, fvars):
params[name].value = params[name].from_internal(val)
else:
for name, val in zip(self.result.var_names, fvars):
params[name].value = val
params.update_constraints()
self.result.nfev += 1
out = self.userfcn(params, *self.userargs, **self.userkws)
if callable(self.iter_cb):
abort = self.iter_cb(params, self.result.nfev, out,
*self.userargs, **self.userkws)
self._abort = self._abort or abort
if self._abort:
self.result.residual = out
self.result.aborted = True
self.result.message = "Fit aborted by user callback. Could not estimate error-bars."
self.result.success = False
raise AbortFitException("fit aborted by user.")
else:
return _nan_policy(np.asarray(out).ravel(),
nan_policy=self.nan_policy)
def __jacobian(self, fvars):
"""Return analytical jacobian to be used with Levenberg-Marquardt.
modified 02-01-2012 by Glenn Jones, Aberystwyth University
modified 06-29-2015 M Newville to apply gradient scaling for
bounded variables (thanks to JJ Helmus, N Mayorov)
"""
pars = self.result.params
grad_scale = ones_like(fvars)
for ivar, name in enumerate(self.result.var_names):
val = fvars[ivar]
pars[name].value = pars[name].from_internal(val)
grad_scale[ivar] = pars[name].scale_gradient(val)
pars.update_constraints()
# compute the jacobian for "internal" unbounded variables,
# then rescale for bounded "external" variables.
jac = self.jacfcn(pars, *self.userargs, **self.userkws)
jac = _nan_policy(jac, nan_policy=self.nan_policy)
if self.col_deriv:
jac = (jac.transpose()*grad_scale).transpose()
else:
jac *= grad_scale
return jac
def penalty(self, fvars):
"""Penalty function for scalar minimizers.
Parameters
----------
fvars : numpy.ndarray
Array of values for the variable parameters.
Returns
-------
r : float
The evaluated user-supplied objective function.
If the objective function is an array of size greater than 1,
use the scalar returned by `self.reduce_fcn`. This defaults
to sum-of-squares, but can be replaced by other options.
"""
if self.result.method == 'brute':
apply_bounds_transformation = False
else:
apply_bounds_transformation = True
r = self.__residual(fvars, apply_bounds_transformation)
if isinstance(r, ndarray) and r.size > 1:
r = self.reduce_fcn(r)
if isinstance(r, ndarray) and r.size > 1:
r = r.sum()
return r
def prepare_fit(self, params=None):
"""Prepare parameters for fitting.
Prepares and initializes model and Parameters for subsequent
fitting. This routine prepares the conversion of :class:`Parameters`
into fit variables, organizes parameter bounds, and parses, "compiles"
and checks constrain expressions. The method also creates and returns
a new instance of a :class:`MinimizerResult` object that contains the
copy of the Parameters that will actually be varied in the fit.
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Contains the Parameters for the model; if None, then the
Parameters used to initialize the Minimizer object are used.
Returns
-------
:class:`MinimizerResult`
Notes
-----
This method is called directly by the fitting methods, and it is
generally not necessary to call this function explicitly.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
"""
# determine which parameters are actually variables
# and which are defined expressions.
self.result = MinimizerResult()
result = self.result
if params is not None:
self.params = params
if isinstance(self.params, Parameters):
result.params = deepcopy(self.params)
elif isinstance(self.params, (list, tuple)):
result.params = Parameters()
for par in self.params:
if not isinstance(par, Parameter):
raise MinimizerException(self._err_nonparam)
else:
result.params[par.name] = par
elif self.params is None:
raise MinimizerException(self._err_nonparam)
# determine which parameters are actually variables
# and which are defined expressions.
result.var_names = [] # note that this *does* belong to self...
result.init_vals = []
result.params.update_constraints()
result.nfev = 0
result.errorbars = False
result.aborted = False
for name, par in self.result.params.items():
par.stderr = None
par.correl = None
if par.expr is not None:
par.vary = False
if par.vary:
result.var_names.append(name)
result.init_vals.append(par.setup_bounds())
par.init_value = par.value
if par.name is None:
par.name = name
result.nvarys = len(result.var_names)
result.init_values = {n: v for n, v in zip(result.var_names,
result.init_vals)}
# set up reduce function for scalar minimizers
# 1. user supplied callable
# 2. string starting with 'neglogc' or 'negent'
# 3. sum of squares
if not callable(self.reduce_fcn):
if isinstance(self.reduce_fcn, six.string_types):
if self.reduce_fcn.lower().startswith('neglogc'):
self.reduce_fcn = reduce_cauchylogpdf
elif self.reduce_fcn.lower().startswith('negent'):
self.reduce_fcn = reduce_negentropy
if self.reduce_fcn is None:
self.reduce_fcn = reduce_chisquare
return result
def unprepare_fit(self):
"""Clean fit state, so that subsequent fits need to call prepare_fit().
removes AST compilations of constraint expressions.
"""
pass
def _calculate_covariance_matrix(self, fvars):
"""Calculate the covariance matrix.
The `numdiftoools` package is used to estimate the Hessian matrix, and
the covariance matrix is calculated as:
.. math::
cov_x = inverse(Hessian) * 2.0
Parameters
----------
fvars : numpy.ndarray
Array of the optimal internal, freely variable parameters.
Returns
-------
cov_x : numpy.ndarray or None
Covariance matrix if successful, otherwise None.
"""
warnings.filterwarnings(action="ignore", module="scipy",
message="^internal gelsd")
nfev = deepcopy(self.result.nfev)
try:
Hfun = ndt.Hessian(self.penalty)
hessian_ndt = Hfun(fvars)
cov_x = inv(hessian_ndt) * 2.0
except (LinAlgError, ValueError):
return None
finally:
self.result.nfev = nfev
return cov_x
def _int2ext_cov_x(self, cov_int, fvars):
"""Transform covariance matrix to external parameter space.
It makes use of the gradient scaling according to the MINUIT recipe:
cov_ext = np.dot(grad.T, grad) * cov_int
Parameters
----------
cov_int : numpy.ndarray
Covariance matrix in the internal parameter space.
fvars : numpy.ndarray
Array of the optimal internal, freely variable, parameter values.
Returns
-------
cov_ext : numpy.ndarray
Covariance matrix, transformed to external parameter space.
"""
g = [self.result.params[name].scale_gradient(fvars[i]) for i, name in
enumerate(self.result.var_names)]
grad2d = np.atleast_2d(g)
grad = np.dot(grad2d.T, grad2d)
cov_ext = cov_int * grad
return cov_ext
def _calculate_uncertainties_correlations(self):
"""Calculate parameter uncertainties and correlations."""
self.result.errorbars = True
if self.scale_covar:
self.result.covar *= self.result.redchi
vbest = np.atleast_1d([self.result.params[name].value for name in
self.result.var_names])
has_expr = False
for par in self.result.params.values():
par.stderr, par.correl = 0, None
has_expr = has_expr or par.expr is not None
for ivar, name in enumerate(self.result.var_names):
par = self.result.params[name]
par.stderr = sqrt(self.result.covar[ivar, ivar])
par.correl = {}
try:
self.result.errorbars = self.result.errorbars and (par.stderr > 0.0)
for jvar, varn2 in enumerate(self.result.var_names):
if jvar != ivar:
par.correl[varn2] = (self.result.covar[ivar, jvar] /
(par.stderr * sqrt(self.result.covar[jvar, jvar])))
except ZeroDivisionError:
self.result.errorbars = False
if has_expr:
try:
uvars = uncertainties.correlated_values(vbest, self.result.covar)
except (LinAlgError, ValueError):
uvars = None
# for uncertainties on constrained parameters, use the calculated
# "correlated_values", evaluate the uncertainties on the constrained
# parameters and reset the Parameters to best-fit value
if uvars is not None:
for par in self.result.params.values():
eval_stderr(par, uvars, self.result.var_names, self.result.params)
# restore nominal values
for v, nam in zip(uvars, self.result.var_names):
self.result.params[nam].value = v.nominal_value
def scalar_minimize(self, method='Nelder-Mead', params=None, **kws):
"""Scalar minimization using :scipydoc:`optimize.minimize`.
Perform fit with any of the scalar minimization algorithms supported by
:scipydoc:`optimize.minimize`. Default argument values are:
+-------------------------+-----------------+-----------------------------------------------------+
| :meth:`scalar_minimize` | Default Value | Description |
| arg | | |
+=========================+=================+=====================================================+
| method | ``Nelder-Mead`` | fitting method |
+-------------------------+-----------------+-----------------------------------------------------+
| tol | 1.e-7 | fitting and parameter tolerance |
+-------------------------+-----------------+-----------------------------------------------------+
| hess | None | Hessian of objective function |
+-------------------------+-----------------+-----------------------------------------------------+
Parameters
----------
method : str, optional
Name of the fitting method to use. One of:
- 'Nelder-Mead' (default)
- 'L-BFGS-B'
- 'Powell'
- 'CG'
- 'Newton-CG'
- 'COBYLA'
- 'BFGS'
- 'TNC'
- 'trust-ncg'
- 'trust-exact' (SciPy >= 1.0)
- 'trust-krylov' (SciPy >= 1.0)
- 'trust-constr' (SciPy >= 1.1)
- 'dogleg'
- 'SLSQP'
- 'differential_evolution'
params : :class:`~lmfit.parameter.Parameters`, optional
Parameters to use as starting point.
**kws : dict, optional
Minimizer options pass to :scipydoc:`optimize.minimize`.
Returns
-------
:class:`MinimizerResult`
Object containing the optimized parameter and several
goodness-of-fit statistics.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
Notes
-----
If the objective function returns a NumPy array instead
of the expected scalar, the sum of squares of the array
will be used.
Note that bounds and constraints can be set on Parameters
for any of these methods, so are not supported separately
for those designed to use bounds. However, if you use the
differential_evolution method you must specify finite
(min, max) for each varying Parameter.
"""
result = self.prepare_fit(params=params)
result.method = method
variables = result.init_vals
params = result.params
fmin_kws = dict(method=method,
options={'maxiter': 1000 * (len(variables) + 1)})
fmin_kws.update(self.kws)
fmin_kws.update(kws)
# hess supported only in some methods
if 'hess' in fmin_kws and method not in ('Newton-CG',
'dogleg', 'trust-ncg'):
fmin_kws.pop('hess')
# jac supported only in some methods (and Dfun could be used...)
if 'jac' not in fmin_kws and fmin_kws.get('Dfun', None) is not None:
self.jacfcn = fmin_kws.pop('jac')
fmin_kws['jac'] = self.__jacobian
if 'jac' in fmin_kws and method not in ('CG', 'BFGS', 'Newton-CG',
'dogleg', 'trust-ncg'):
self.jacfcn = None
fmin_kws.pop('jac')
if method == 'differential_evolution':
for par in params.values():
if (par.vary and
not (np.isfinite(par.min) and np.isfinite(par.max))):
raise ValueError('differential_evolution requires finite '
'bound for all varying parameters')
_bounds = [(-np.pi / 2., np.pi / 2.)] * len(variables)
kwargs = dict(args=(), strategy='best1bin', maxiter=None,
popsize=15, tol=0.01, mutation=(0.5, 1),
recombination=0.7, seed=None, callback=None,
disp=False, polish=True, init='latinhypercube')
for k, v in fmin_kws.items():
if k in kwargs:
kwargs[k] = v
try:
ret = differential_evolution(self.penalty, _bounds, **kwargs)
except AbortFitException:
pass
else:
try:
ret = scipy_minimize(self.penalty, variables, **fmin_kws)
except AbortFitException:
pass
if not result.aborted:
if isinstance(ret, dict):
for attr, value in ret.items():
setattr(result, attr, value)
else:
for attr in dir(ret):
if not attr.startswith('_'):
setattr(result, attr, getattr(ret, attr))
result.x = np.atleast_1d(result.x)
result.residual = self.__residual(result.x)
result.nfev -= 1
else:
pass
result._calculate_statistics()
# calculate the cov_x and estimate uncertanties/correlations
if (self.calc_covar and HAS_NUMDIFFTOOLS and
len(result.residual) > len(result.var_names)):
_covar_ndt = self._calculate_covariance_matrix(result.x)
if _covar_ndt is not None:
result.covar = self._int2ext_cov_x(_covar_ndt, result.x)
self._calculate_uncertainties_correlations()
return result
def emcee(self, params=None, steps=1000, nwalkers=100, burn=0, thin=1,
ntemps=1, pos=None, reuse_sampler=False, workers=1,
float_behavior='posterior', is_weighted=True, seed=None, progress=True):
r"""Bayesian sampling of the posterior distribution using `emcee`.
Bayesian sampling of the posterior distribution for the parameters
using the `emcee` Markov Chain Monte Carlo package. The method assumes
that the prior is Uniform. You need to have `emcee` installed to use
this method.
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Parameters to use as starting point. If this is not specified
then the Parameters used to initialize the Minimizer object are
used.
steps : int, optional
How many samples you would like to draw from the posterior
distribution for each of the walkers?
nwalkers : int, optional
Should be set so :math:`nwalkers >> nvarys`, where `nvarys` are
the number of parameters being varied during the fit.
"Walkers are the members of the ensemble. They are almost like
separate Metropolis-Hastings chains but, of course, the proposal
distribution for a given walker depends on the positions of all
the other walkers in the ensemble." - from the `emcee` webpage.
burn : int, optional
Discard this many samples from the start of the sampling regime.
thin : int, optional
Only accept 1 in every `thin` samples.
ntemps : int, optional
If `ntemps > 1` perform a Parallel Tempering.
pos : numpy.ndarray, optional
Specify the initial positions for the sampler. If `ntemps == 1`
then `pos.shape` should be `(nwalkers, nvarys)`. Otherwise,
`(ntemps, nwalkers, nvarys)`. You can also initialise using a
previous chain that had the same `ntemps`, `nwalkers` and
`nvarys`. Note that `nvarys` may be one larger than you expect it
to be if your `userfcn` returns an array and `is_weighted is
False`.
reuse_sampler : bool, optional
If you have already run `emcee` on a given `Minimizer` object then
it possesses an internal ``sampler`` attribute. You can continue to
draw from the same sampler (retaining the chain history) if you set
this option to True. Otherwise a new sampler is created. The
`nwalkers`, `ntemps`, `pos`, and `params` keywords are ignored with
this option.
**Important**: the Parameters used to create the sampler must not
change in-between calls to `emcee`. Alteration of Parameters
would include changed ``min``, ``max``, ``vary`` and ``expr``
attributes. This may happen, for example, if you use an altered
Parameters object and call the `minimize` method in-between calls
to `emcee`.
workers : Pool-like or int, optional
For parallelization of sampling. It can be any Pool-like object
with a map method that follows the same calling sequence as the
built-in `map` function. If int is given as the argument, then a
multiprocessing-based pool is spawned internally with the
corresponding number of parallel processes. 'mpi4py'-based
parallelization and 'joblib'-based parallelization pools can also
be used here. **Note**: because of multiprocessing overhead it may
only be worth parallelising if the objective function is expensive
to calculate, or if there are a large number of objective
evaluations per step (`ntemps * nwalkers * nvarys`).
float_behavior : str, optional
Specifies meaning of the objective function output if it returns a
float. One of:
- 'posterior' - objective function returns a log-posterior
probability
- 'chi2' - objective function returns :math:`\chi^2`
See Notes for further details.
is_weighted : bool, optional
Has your objective function been weighted by measurement
uncertainties? If `is_weighted is True` then your objective
function is assumed to return residuals that have been divided by
the true measurement uncertainty `(data - model) / sigma`. If
`is_weighted is False` then the objective function is assumed to
return unweighted residuals, `data - model`. In this case `emcee`
will employ a positive measurement uncertainty during the sampling.
This measurement uncertainty will be present in the output params
and output chain with the name `__lnsigma`. A side effect of this
is that you cannot use this parameter name yourself.
**Important** this parameter only has any effect if your objective
function returns an array. If your objective function returns a
float, then this parameter is ignored. See Notes for more details.
seed : int or `numpy.random.RandomState`, optional
If `seed` is an int, a new `numpy.random.RandomState` instance is
used, seeded with `seed`.
If `seed` is already a `numpy.random.RandomState` instance, then
that `numpy.random.RandomState` instance is used.
Specify `seed` for repeatable minimizations.
Returns
-------
:class:`MinimizerResult`
MinimizerResult object containing updated params, statistics,
etc. The updated params represent the median (50th percentile) of
all the samples, whilst the parameter uncertainties are half of the
difference between the 15.87 and 84.13 percentiles.
The `MinimizerResult` also contains the ``chain``, ``flatchain``
and ``lnprob`` attributes. The ``chain`` and ``flatchain``
attributes contain the samples and have the shape
`(nwalkers, (steps - burn) // thin, nvarys)` or
`(ntemps, nwalkers, (steps - burn) // thin, nvarys)`,
depending on whether Parallel tempering was used or not.
`nvarys` is the number of parameters that are allowed to vary.
The ``flatchain`` attribute is a `pandas.DataFrame` of the
flattened chain, `chain.reshape(-1, nvarys)`. To access flattened
chain values for a particular parameter use
`result.flatchain[parname]`. The ``lnprob`` attribute contains the
log probability for each sample in ``chain``. The sample with the
highest probability corresponds to the maximum likelihood estimate.
Notes
-----
This method samples the posterior distribution of the parameters using
Markov Chain Monte Carlo. To do so it needs to calculate the
log-posterior probability of the model parameters, `F`, given the data,
`D`, :math:`\ln p(F_{true} | D)`. This 'posterior probability' is
calculated as:
.. math::
\ln p(F_{true} | D) \propto \ln p(D | F_{true}) + \ln p(F_{true})
where :math:`\ln p(D | F_{true})` is the 'log-likelihood' and
:math:`\ln p(F_{true})` is the 'log-prior'. The default log-prior
encodes prior information already known about the model. This method
assumes that the log-prior probability is `-numpy.inf` (impossible) if
the one of the parameters is outside its limits. The log-prior probability
term is zero if all the parameters are inside their bounds (known as a
uniform prior). The log-likelihood function is given by [1]_:
.. math::
\ln p(D|F_{true}) = -\frac{1}{2}\sum_n \left[\frac{(g_n(F_{true}) - D_n)^2}{s_n^2}+\ln (2\pi s_n^2)\right]
The first summand in the square brackets represents the residual for a
given datapoint (:math:`g` being the generative model, :math:`D_n` the
data and :math:`s_n` the standard deviation, or measurement
uncertainty, of the datapoint). This term represents :math:`\chi^2`
when summed over all data points.
Ideally the objective function used to create `lmfit.Minimizer` should
return the log-posterior probability, :math:`\ln p(F_{true} | D)`.
However, since the in-built log-prior term is zero, the objective
function can also just return the log-likelihood, unless you wish to
create a non-uniform prior.
If a float value is returned by the objective function then this value
is assumed by default to be the log-posterior probability, i.e.
`float_behavior is 'posterior'`. If your objective function returns
:math:`\chi^2`, then you should use a value of `'chi2'` for
`float_behavior`. `emcee` will then multiply your :math:`\chi^2` value
by -0.5 to obtain the posterior probability.
However, the default behaviour of many objective functions is to return
a vector of (possibly weighted) residuals. Therefore, if your objective
function returns a vector, `res`, then the vector is assumed to contain
the residuals. If `is_weighted is True` then your residuals are assumed
to be correctly weighted by the standard deviation (measurement
uncertainty) of the data points (`res = (data - model) / sigma`) and
the log-likelihood (and log-posterior probability) is calculated as:
`-0.5 * numpy.sum(res**2)`.
This ignores the second summand in the square brackets. Consequently,
in order to calculate a fully correct log-posterior probability value
your objective function should return a single value. If
`is_weighted is False` then the data uncertainty, `s_n`, will be
treated as a nuisance parameter and will be marginalized out. This is
achieved by employing a strictly positive uncertainty
(homoscedasticity) for each data point, :math:`s_n = \exp(\_\_lnsigma)`.
`__lnsigma` will be present in `MinimizerResult.params`, as well as
`Minimizer.chain`, `nvarys` will also be increased by one.
References
----------
.. [1] http://dan.iel.fm/emcee/current/user/line/
"""
if not HAS_EMCEE:
raise NotImplementedError('You must have emcee to use'
' the emcee method')
tparams = params
# if you're reusing the sampler then ntemps, nwalkers have to be
# determined from the previous sampling
if reuse_sampler:
if not hasattr(self, 'sampler') or not hasattr(self, '_lastpos'):
raise ValueError("You wanted to use an existing sampler, but"
"it hasn't been created yet")
if len(self._lastpos.shape) == 2:
ntemps = 1
nwalkers = self._lastpos.shape[0]
elif len(self._lastpos.shape) == 3:
ntemps = self._lastpos.shape[0]
nwalkers = self._lastpos.shape[1]
tparams = None
result = self.prepare_fit(params=tparams)
result.method = 'emcee'
params = result.params
# check if the userfcn returns a vector of residuals
out = self.userfcn(params, *self.userargs, **self.userkws)
out = np.asarray(out).ravel()
if out.size > 1 and is_weighted is False:
# we need to marginalise over a constant data uncertainty
if '__lnsigma' not in params:
# __lnsigma should already be in params if is_weighted was
# previously set to True.
params.add('__lnsigma', value=0.01, min=-np.inf, max=np.inf, vary=True)
# have to re-prepare the fit
result = self.prepare_fit(params)
params = result.params
# Removing internal parameter scaling. We could possibly keep it,
# but I don't know how this affects the emcee sampling.
bounds = []
var_arr = np.zeros(len(result.var_names))
i = 0
for par in params:
param = params[par]
if param.expr is not None:
param.vary = False
if param.vary:
var_arr[i] = param.value
i += 1
else:
# don't want to append bounds if they're not being varied.
continue
param.from_internal = lambda val: val
lb, ub = param.min, param.max
if lb is None or lb is np.nan:
lb = -np.inf
if ub is None or ub is np.nan:
ub = np.inf
bounds.append((lb, ub))
bounds = np.array(bounds)
self.nvarys = len(result.var_names)
# set up multiprocessing options for the samplers
auto_pool = None
sampler_kwargs = {}
if isinstance(workers, int) and workers > 1:
auto_pool = multiprocessing.Pool(workers)
sampler_kwargs['pool'] = auto_pool
elif hasattr(workers, 'map'):
sampler_kwargs['pool'] = workers
# function arguments for the log-probability functions
# these values are sent to the log-probability functions by the sampler.
lnprob_args = (self.userfcn, params, result.var_names, bounds)
lnprob_kwargs = {'is_weighted': is_weighted,
'float_behavior': float_behavior,
'userargs': self.userargs,
'userkws': self.userkws,
'nan_policy': self.nan_policy}
if ntemps > 1:
# the prior and likelihood function args and kwargs are the same
sampler_kwargs['loglargs'] = lnprob_args
sampler_kwargs['loglkwargs'] = lnprob_kwargs
sampler_kwargs['logpargs'] = (bounds,)
else:
sampler_kwargs['args'] = lnprob_args
sampler_kwargs['kwargs'] = lnprob_kwargs
# set up the random number generator
rng = _make_random_gen(seed)
# now initialise the samplers
if reuse_sampler:
if auto_pool is not None:
self.sampler.pool = auto_pool
p0 = self._lastpos
if p0.shape[-1] != self.nvarys:
raise ValueError("You cannot reuse the sampler if the number"
"of varying parameters has changed")
elif ntemps > 1:
# Parallel Tempering
# jitter the starting position by scaled Gaussian noise
p0 = 1 + rng.randn(ntemps, nwalkers, self.nvarys) * 1.e-4
p0 *= var_arr
self.sampler = emcee.PTSampler(ntemps, nwalkers, self.nvarys,
_lnpost, _lnprior, **sampler_kwargs)
else:
p0 = 1 + rng.randn(nwalkers, self.nvarys) * 1.e-4
p0 *= var_arr
self.sampler = emcee.EnsembleSampler(nwalkers, self.nvarys,
_lnpost, **sampler_kwargs)
# user supplies an initialisation position for the chain
# If you try to run the sampler with p0 of a wrong size then you'll get
# a ValueError. Note, you can't initialise with a position if you are
# reusing the sampler.
if pos is not None and not reuse_sampler:
tpos = np.asfarray(pos)
if p0.shape == tpos.shape:
pass
# trying to initialise with a previous chain
elif tpos.shape[0::2] == (nwalkers, self.nvarys):
tpos = tpos[:, -1, :]
# initialising with a PTsampler chain.
elif ntemps > 1 and tpos.ndim == 4:
tpos_shape = list(tpos.shape)
tpos_shape.pop(2)
if tpos_shape == (ntemps, nwalkers, self.nvarys):
tpos = tpos[..., -1, :]
else:
raise ValueError('pos should have shape (nwalkers, nvarys)'
'or (ntemps, nwalkers, nvarys) if ntemps > 1')
p0 = tpos
# if you specified a seed then you also need to seed the sampler
if seed is not None:
self.sampler.random_state = rng.get_state()
# now do a production run, sampling all the time
if EMCEE_VERSION >= 3:
output = self.sampler.run_mcmc(p0, steps, progress=progress)
self._lastpos = output.coords
else:
output = self.sampler.run_mcmc(p0, steps)
self._lastpos = output[0]
# discard the burn samples and thin
chain = self.sampler.chain[..., burn::thin, :]
lnprobability = self.sampler.lnprobability[..., burn::thin]
# take the zero'th PTsampler temperature for the parameter estimators
if ntemps > 1:
flatchain = chain[0, ...].reshape((-1, self.nvarys))
else:
flatchain = chain.reshape((-1, self.nvarys))
quantiles = np.percentile(flatchain, [15.87, 50, 84.13], axis=0)
for i, var_name in enumerate(result.var_names):
std_l, median, std_u = quantiles[:, i]
params[var_name].value = median
params[var_name].stderr = 0.5 * (std_u - std_l)
params[var_name].correl = {}
params.update_constraints()
# work out correlation coefficients
corrcoefs = np.corrcoef(flatchain.T)
for i, var_name in enumerate(result.var_names):
for j, var_name2 in enumerate(result.var_names):
if i != j:
result.params[var_name].correl[var_name2] = corrcoefs[i, j]
result.chain = np.copy(chain)
result.lnprob = np.copy(lnprobability)
result.errorbars = True
result.nvarys = len(result.var_names)
result.nfev = ntemps*nwalkers*steps
# Calculate the residual with the "best fit" parameters
out = self.userfcn(params, *self.userargs, **self.userkws)
result.residual = _nan_policy(out, nan_policy=self.nan_policy, handle_inf=False)
# If uncertainty was automatically estimated, weight the residual properly
if (not is_weighted) and (result.residual.size > 1):
if '__lnsigma' in params:
result.residual = result.residual/np.exp(params['__lnsigma'].value)
# Calculate statistics for the two standard cases:
if isinstance(result.residual, ndarray) or (float_behavior == 'chi2'):
result._calculate_statistics()
# Handle special case unique to emcee:
# This should eventually be moved into result._calculate_statistics.
elif float_behavior == 'posterior':
result.ndata = 1
result.nfree = 1
# assuming prior prob = 1, this is true
_neg2_log_likel = -2*result.residual
# assumes that residual is properly weighted
result.chisqr = np.exp(_neg2_log_likel)
result.redchi = result.chisqr / result.nfree
result.aic = _neg2_log_likel + 2 * result.nvarys
result.bic = _neg2_log_likel + np.log(result.ndata) * result.nvarys
if auto_pool is not None:
auto_pool.terminate()
return result
def least_squares(self, params=None, **kws):
"""Least-squares minimization using :scipydoc:`optimize.least_squares`.
This method wraps :scipydoc:`optimize.least_squares`, which has inbuilt
support for bounds and robust loss functions. By default it uses the
Trust Region Reflective algorithm with a linear loss function (i.e.,
the standard least-squares problem).
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Parameters to use as starting point.
**kws : dict, optional
Minimizer options to pass to :scipydoc:`optimize.least_squares`.
Returns
-------
:class:`MinimizerResult`
Object containing the optimized parameter and several
goodness-of-fit statistics.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
"""
result = self.prepare_fit(params)
result.method = 'least_squares'
replace_none = lambda x, sign: sign*np.inf if x is None else x
start_vals, lower_bounds, upper_bounds = [], [], []
for vname in result.var_names:
par = self.params[vname]
start_vals.append(par.value)
lower_bounds.append(replace_none(par.min, -1))
upper_bounds.append(replace_none(par.max, 1))
try:
ret = least_squares(self.__residual, start_vals,
bounds=(lower_bounds, upper_bounds),
kwargs=dict(apply_bounds_transformation=False),
**kws)
except AbortFitException:
pass
if not result.aborted:
for attr in ret:
setattr(result, attr, ret[attr])
result.x = np.atleast_1d(result.x)
result.residual = ret.fun
else:
pass
result._calculate_statistics()
# calculate the cov_x and estimate uncertainties/correlations
try:
hess = np.matmul(ret.jac.T, ret.jac)
result.covar = np.linalg.inv(hess)
self._calculate_uncertainties_correlations()
except LinAlgError:
pass
return result
def leastsq(self, params=None, **kws):
"""Use Levenberg-Marquardt minimization to perform a fit.
It assumes that the input Parameters have been initialized, and
a function to minimize has been properly set up.
When possible, this calculates the estimated uncertainties and
variable correlations from the covariance matrix.
This method calls :scipydoc:`optimize.leastsq`.
By default, numerical derivatives are used, and the following
arguments are set:
+------------------+----------------+------------------------------------------------------------+
| :meth:`leastsq` | Default Value | Description |
| arg | | |
+==================+================+============================================================+
| xtol | 1.e-7 | Relative error in the approximate solution |
+------------------+----------------+------------------------------------------------------------+
| ftol | 1.e-7 | Relative error in the desired sum of squares |
+------------------+----------------+------------------------------------------------------------+
| maxfev | 2000*(nvar+1) | Maximum number of function calls (nvar= # of variables) |
+------------------+----------------+------------------------------------------------------------+
| Dfun | None | Function to call for Jacobian calculation |
+------------------+----------------+------------------------------------------------------------+
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Parameters to use as starting point.
**kws : dict, optional
Minimizer options to pass to :scipydoc:`optimize.leastsq`.
Returns
-------
:class:`MinimizerResult`
Object containing the optimized parameter
and several goodness-of-fit statistics.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
"""
result = self.prepare_fit(params=params)
result.method = 'leastsq'
result.nfev -= 2 # correct for "pre-fit" initialization/checks
variables = result.init_vals
nvars = len(variables)
lskws = dict(full_output=1, xtol=1.e-7, ftol=1.e-7, col_deriv=False,
gtol=1.e-7, maxfev=2000*(nvars+1), Dfun=None)
lskws.update(self.kws)
lskws.update(kws)
self.col_deriv = False
if lskws['Dfun'] is not None:
self.jacfcn = lskws['Dfun']
self.col_deriv = lskws['col_deriv']
lskws['Dfun'] = self.__jacobian
# suppress runtime warnings during fit and error analysis
orig_warn_settings = np.geterr()
np.seterr(all='ignore')
try:
lsout = scipy_leastsq(self.__residual, variables, **lskws)
_best, _cov, infodict, errmsg, ier = lsout
result.residual = infodict['fvec']
except AbortFitException:
pass
result._calculate_statistics()
if result.aborted:
return result
result.ier = ier
result.lmdif_message = errmsg
result.success = ier in [1, 2, 3, 4]
if ier in {1, 2, 3}:
result.message = 'Fit succeeded.'
elif ier == 0:
result.message = ('Invalid Input Parameters. I.e. more variables '
'than data points given, tolerance < 0.0, or '
'no data provided.')
elif ier == 4:
result.message = 'One or more variable did not affect the fit.'
elif ier == 5:
result.message = self._err_maxfev % lskws['maxfev']
else:
result.message = 'Tolerance seems to be too small.'
# self.errorbars = error bars were successfully estimated
result.errorbars = (_cov is not None)
if result.errorbars:
# transform the covariance matrix to "external" parameter space
result.covar = self._int2ext_cov_x(_cov, _best)
# calculate parameter uncertainties and correlations
self._calculate_uncertainties_correlations()
else:
result.message = '%s Could not estimate error-bars.' % result.message
np.seterr(**orig_warn_settings)
return result
def basinhopping(self, params=None, **kws):
"""Use the `basinhopping` algorithm to find the global minimum of a function.
This method calls :scipydoc:`optimize.basinhopping` using the default
arguments. The default minimizer is `BFGS`, but since lmfit supports
parameter bounds for all minimizers, the user can choose any of the
solvers present in :scipydoc:`optimize.minimize`.
Parameters
----------
params : :class:`~lmfit.parameter.Parameters` object, optional
Contains the Parameters for the model. If None, then the
Parameters used to initialize the Minimizer object are used.
Returns
-------
:class:`MinimizerResult`
Object containing the optimization results from the basinhopping
algorithm.
.. versionadded:: 0.9.10
"""
result = self.prepare_fit(params=params)
result.method = 'basinhopping'
basinhopping_kws = dict(niter=100, T=1.0, stepsize=0.5,
minimizer_kwargs={}, take_step=None,
accept_test=None, callback=None, interval=50,
disp=False, niter_success=None, seed=None)
basinhopping_kws.update(self.kws)
basinhopping_kws.update(kws)
# FIXME - remove after requirement for scipy >= 0.19
major, minor, micro = scipy_version.split('.', 2)
if int(major) < 1 and int(minor) < 19:
_ = basinhopping_kws.pop('seed')
print("Warning: basinhopping doesn't support argument 'seed' for "
"scipy versions below 0.19!")
x0 = result.init_vals
try:
ret = scipy_basinhopping(self.penalty, x0, **basinhopping_kws)
except AbortFitException:
pass
if not result.aborted:
result.message = ret.message
result.residual = self.__residual(ret.x)
result.nfev -= 1
else:
pass
result._calculate_statistics()
# calculate the cov_x and estimate uncertanties/correlations
if (self.calc_covar and HAS_NUMDIFFTOOLS and
len(result.residual) > len(result.var_names)):
_covar_ndt = self._calculate_covariance_matrix(ret.x)
if _covar_ndt is not None:
result.covar = self._int2ext_cov_x(_covar_ndt, ret.x)
self._calculate_uncertainties_correlations()
return result
def brute(self, params=None, Ns=20, keep=50):
"""Use the `brute` method to find the global minimum of a function.
The following parameters are passed to :scipydoc:`optimize.brute`
and cannot be changed:
+-------------------+-------+----------------------------------------+
| :meth:`brute` arg | Value | Description |
+===================+=======+========================================+
| full_output | 1 | Return the evaluation grid and |
| | | the objective function's values on it. |
+-------------------+-------+----------------------------------------+
| finish | None | No "polishing" function is to be used |
| | | after the grid search. |
+-------------------+-------+----------------------------------------+
| disp | False | Do not print convergence messages |
| | | (when finish is not None). |
+-------------------+-------+----------------------------------------+
It assumes that the input Parameters have been initialized, and a
function to minimize has been properly set up.
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Contains the Parameters for the model. If None, then the
Parameters used to initialize the Minimizer object are used.
Ns : int, optional
Number of grid points along the axes, if not otherwise specified
(see Notes).
keep : int, optional
Number of best candidates from the brute force method that are
stored in the :attr:`candidates` attribute. If 'all', then all grid
points from :scipydoc:`optimize.brute` are stored as candidates.
Returns
-------
:class:`MinimizerResult`
Object containing the parameters from the brute force method.
The return values (`x0`, `fval`, `grid`, `Jout`) from
:scipydoc:`optimize.brute` are stored as `brute_<parname>` attributes.
The `MinimizerResult` also contains the `candidates` attribute and
`show_candidates()` method. The `candidates` attribute contains the
parameters and chisqr from the brute force method as a namedtuple,
('Candidate', ['params', 'score']), sorted on the (lowest) chisqr
value. To access the values for a particular candidate one can use
`result.candidate[#].params` or `result.candidate[#].score`, where
a lower # represents a better candidate. The `show_candidates(#)`
uses the :meth:`pretty_print` method to show a specific candidate-#
or all candidates when no number is specified.
.. versionadded:: 0.9.6
Notes
-----
The :meth:`brute` method evalutes the function at each point of a
multidimensional grid of points. The grid points are generated from the
parameter ranges using `Ns` and (optional) `brute_step`.
The implementation in :scipydoc:`optimize.brute` requires finite bounds
and the `range` is specified as a two-tuple `(min, max)` or slice-object
`(min, max, brute_step)`. A slice-object is used directly, whereas a
two-tuple is converted to a slice object that interpolates `Ns` points
from `min` to `max`, inclusive.
In addition, the :meth:`brute` method in lmfit, handles three other
scenarios given below with their respective slice-object:
- lower bound (:attr:`min`) and :attr:`brute_step` are specified:
range = (`min`, `min` + `Ns` * `brute_step`, `brute_step`).
- upper bound (:attr:`max`) and :attr:`brute_step` are specified:
range = (`max` - `Ns` * `brute_step`, `max`, `brute_step`).
- numerical value (:attr:`value`) and :attr:`brute_step` are specified:
range = (`value` - (`Ns`//2) * `brute_step`, `value` +
(`Ns`//2) * `brute_step`, `brute_step`).
"""
result = self.prepare_fit(params=params)
result.method = 'brute'
result.nfev -= 1 # correct for "pre-fit" initialization/checks
brute_kws = dict(full_output=1, finish=None, disp=False)
varying = np.asarray([par.vary for par in self.params.values()])
replace_none = lambda x, sign: sign*np.inf if x is None else x
lower_bounds = np.asarray([replace_none(i.min, -1) for i in
self.params.values()])[varying]
upper_bounds = np.asarray([replace_none(i.max, 1) for i in
self.params.values()])[varying]
value = np.asarray([i.value for i in self.params.values()])[varying]
stepsize = np.asarray([i.brute_step for i in self.params.values()])[varying]
ranges = []
for i, step in enumerate(stepsize):
if np.all(np.isfinite([lower_bounds[i], upper_bounds[i]])):
# lower AND upper bounds are specified (brute_step optional)
par_range = ((lower_bounds[i], upper_bounds[i], step)
if step else (lower_bounds[i], upper_bounds[i]))
elif np.isfinite(lower_bounds[i]) and step:
# lower bound AND brute_step are specified
par_range = (lower_bounds[i], lower_bounds[i] + Ns*step, step)
elif np.isfinite(upper_bounds[i]) and step:
# upper bound AND brute_step are specified
par_range = (upper_bounds[i] - Ns*step, upper_bounds[i], step)
elif np.isfinite(value[i]) and step:
# no bounds, but an initial value is specified
par_range = (value[i] - (Ns//2)*step, value[i] + (Ns//2)*step,
step)
else:
raise ValueError('Not enough information provided for the brute '
'force method. Please specify bounds or at '
'least an initial value and brute_step for '
'parameter "{}".'.format(result.var_names[i]))
ranges.append(par_range)
try:
ret = scipy_brute(self.penalty, tuple(ranges), Ns=Ns, **brute_kws)
except AbortFitException:
pass
if not result.aborted:
result.brute_x0 = ret[0]
result.brute_fval = ret[1]
result.brute_grid = ret[2]
result.brute_Jout = ret[3]
# sort the results of brute and populate .candidates attribute
grid_score = ret[3].ravel() # chisqr
grid_points = [par.ravel() for par in ret[2]]
if len(result.var_names) == 1:
grid_result = np.array([res for res in zip(zip(grid_points), grid_score)],
dtype=[('par', 'O'), ('score', 'float64')])
else:
grid_result = np.array([res for res in zip(zip(*grid_points), grid_score)],
dtype=[('par', 'O'), ('score', 'float64')])
grid_result_sorted = grid_result[grid_result.argsort(order='score')]
result.candidates = []
if keep == 'all':
keep_candidates = len(grid_result_sorted)
else:
keep_candidates = min(len(grid_result_sorted), keep)
for data in grid_result_sorted[:keep_candidates]:
pars = deepcopy(self.params)
for i, par in enumerate(result.var_names):
pars[par].value = data[0][i]
result.candidates.append(Candidate(params=pars, score=data[1]))
result.params = result.candidates[0].params
result.residual = self.__residual(result.brute_x0, apply_bounds_transformation=False)
result.nfev -= 1
else:
pass
result._calculate_statistics()
return result
def ampgo(self, params=None, **kws):
"""Find the global minimum of a multivariate function using AMPGO.
AMPGO stands for 'Adaptive Memory Programming for Global Optimization'
and is an efficient algorithm to find the global minimum.
Parameters
----------
params : :class:`~lmfit.parameter.Parameters`, optional
Contains the Parameters for the model. If None, then the
Parameters used to initialize the Minimizer object are used.
**kws : dict, optional
Minimizer options to pass to the ampgo algorithm, the options are
listed below::
local: str (default is 'L-BFGS-B')
Name of the local minimization method. Valid options are:
- 'L-BFGS-B'
- 'Nelder-Mead'
- 'Powell'
- 'TNC'
- 'SLSQP'
local_opts: dict (default is None)
Options to pass to the local minimizer.
maxfunevals: int (default is None)
Maximum number of function evaluations. If None, the optimization will stop
after `totaliter` number of iterations.
totaliter: int (default is 20)
Maximum number of global iterations.
maxiter: int (default is 5)
Maximum number of `Tabu Tunneling` iterations during each global iteration.
glbtol: float (default is 1e-5)
Tolerance whether or not to accept a solution after a tunneling phase.
eps1: float (default is 0.02)
Constant used to define an aspiration value for the objective function during
the Tunneling phase.
eps2: float (default is 0.1)
Perturbation factor used to move away from the latest local minimum at the
start of a Tunneling phase.
tabulistsize: int (default is 5)
Size of the (circular) tabu search list.
tabustrategy: str (default is 'farthest')
Strategy to use when the size of the tabu list exceeds `tabulistsize`. It
can be 'oldest' to drop the oldest point from the tabu list or 'farthest'
to drop the element farthest from the last local minimum found.
disp: bool (default is False)
Set to True to print convergence messages.
Returns
-------
:class:`MinimizerResult`
Object containing the parameters from the ampgo method, with fit
parameters, statistics and such. The return values (`x0`, `fval`,
`eval`, `msg`, `tunnel`) are stored as `ampgo_<parname>` attributes.
.. versionadded:: 0.9.10
Notes
----
The Python implementation was written by Andrea Gavana in 2014
(http://infinity77.net/global_optimization/index.html).
The details of the AMPGO algorithm are described in the paper
"Adaptive Memory Programming for Constrained Global Optimization"
located here:
http://leeds-faculty.colorado.edu/glover/fred%20pubs/416%20-%20AMP%20(TS)%20for%20Constrained%20Global%20Opt%20w%20Lasdon%20et%20al%20.pdf
"""
result = self.prepare_fit(params=params)
ampgo_kws = dict(local='L-BFGS-B', local_opts=None, maxfunevals=None,
totaliter=20, maxiter=5, glbtol=1e-5, eps1=0.02,
eps2=0.1, tabulistsize=5, tabustrategy='farthest',
disp=False)
ampgo_kws.update(self.kws)
ampgo_kws.update(kws)
values = result.init_vals
result.method = "ampgo, with {} as local solver".format(ampgo_kws['local'])
try:
ret = ampgo(self.penalty, values, **ampgo_kws)
except AbortFitException:
pass
if not result.aborted:
result.ampgo_x0 = ret[0]
result.ampgo_fval = ret[1]
result.ampgo_eval = ret[2]
result.ampgo_msg = ret[3]
result.ampgo_tunnel = ret[4]
for i, par in enumerate(result.var_names):
result.params[par].value = result.ampgo_x0[i]
result.residual = self.__residual(result.ampgo_x0)
result.nfev -= 1
else:
pass
result._calculate_statistics()
# calculate the cov_x and estimate uncertanties/correlations
if (self.calc_covar and HAS_NUMDIFFTOOLS and
len(result.residual) > len(result.var_names)):
_covar_ndt = self._calculate_covariance_matrix(result.ampgo_x0)
if _covar_ndt is not None:
result.covar = self._int2ext_cov_x(_covar_ndt, result.ampgo_x0)
self._calculate_uncertainties_correlations()
return result
def minimize(self, method='leastsq', params=None, **kws):
"""Perform the minimization.
Parameters
----------
method : str, optional
Name of the fitting method to use. Valid values are:
- `'leastsq'`: Levenberg-Marquardt (default)
- `'least_squares'`: Least-Squares minimization, using Trust Region Reflective method
- `'differential_evolution'`: differential evolution
- `'brute'`: brute force method
- `'basinhopping'`: basinhopping
- `'ampgo'`: Adaptive Memory Programming for Global Optimization
- '`nelder`': Nelder-Mead
- `'lbfgsb'`: L-BFGS-B
- `'powell'`: Powell
- `'cg'`: Conjugate-Gradient
- `'newton'`: Newton-CG
- `'cobyla'`: Cobyla
- `'bfgs'`: BFGS
- `'tnc'`: Truncated Newton
- `'trust-ncg'`: Newton-CG trust-region
- `'trust-exact'`: nearly exact trust-region (SciPy >= 1.0)
- `'trust-krylov'`: Newton GLTR trust-region (SciPy >= 1.0)
- `'trust-constr'`: trust-region for constrained optimization (SciPy >= 1.1)
- `'dogleg'`: Dog-leg trust-region
- `'slsqp'`: Sequential Linear Squares Programming
- `'emcee'`: Maximum likelihood via Monte-Carlo Markov Chain
In most cases, these methods wrap and use the method with the
same name from `scipy.optimize`, or use
`scipy.optimize.minimize` with the same `method` argument.
Thus '`leastsq`' will use `scipy.optimize.leastsq`, while
'`powell`' will use `scipy.optimize.minimizer(...,
method='powell')`
For more details on the fitting methods please refer to the
`SciPy docs <https://docs.scipy.org/doc/scipy/reference/optimize.html>`__.
params : :class:`~lmfit.parameter.Parameters`, optional
Parameters of the model to use as starting values.
**kws : optional
Additional arguments are passed to the underlying minimization
method.
Returns
-------
:class:`MinimizerResult`
Object containing the optimized parameter and several
goodness-of-fit statistics.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
"""
function = self.leastsq
kwargs = {'params': params}
kwargs.update(self.kws)
kwargs.update(kws)
user_method = method.lower()
if user_method.startswith('leasts'):
function = self.leastsq
elif user_method.startswith('least_s'):
function = self.least_squares
elif user_method == 'brute':
function = self.brute
elif user_method == 'basinhopping':
function = self.basinhopping
elif user_method == 'ampgo':
function = self.ampgo
elif user_method == 'emcee':
function = self.emcee
else:
function = self.scalar_minimize
for key, val in SCALAR_METHODS.items():
if (key.lower().startswith(user_method) or
val.lower().startswith(user_method)):
kwargs['method'] = val
return function(**kwargs)
def _lnprior(theta, bounds):
"""Calculate an improper uniform log-prior probability.
Parameters
----------
theta : sequence
Float parameter values (only those being varied).
bounds : np.ndarray
Lower and upper bounds of parameters that are varying.
Has shape (nvarys, 2).
Returns
-------
lnprob : float
Log prior probability.
"""
if np.any(theta > bounds[:, 1]) or np.any(theta < bounds[:, 0]):
return -np.inf
return 0
def _lnpost(theta, userfcn, params, var_names, bounds, userargs=(),
userkws=None, float_behavior='posterior', is_weighted=True,
nan_policy='raise'):
"""Calculate the log-posterior probability.
See the `Minimizer.emcee` method for more details.
Parameters
----------
theta : sequence
Float parameter values (only those being varied).
userfcn : callable
User objective function.
params : :class:`~lmfit.parameters.Parameters`
The entire set of Parameters.
var_names : list
The names of the parameters that are varying.
bounds : numpy.ndarray
Lower and upper bounds of parameters. Has shape (nvarys, 2).
userargs : tuple, optional
Extra positional arguments required for user objective function.
userkws : dict, optional
Extra keyword arguments required for user objective function.
float_behavior : str, optional
Specifies meaning of objective when it returns a float. One of:
'posterior' - objective function returnins a log-posterior
probability
'chi2' - objective function returns a chi2 value
is_weighted : bool
If `userfcn` returns a vector of residuals then `is_weighted`
specifies if the residuals have been weighted by data uncertainties.
nan_policy : str, optional
Specifies action if `userfcn` returns NaN values. One of:
'raise' - a `ValueError` is raised
'propagate' - the values returned from `userfcn` are un-altered
'omit' - the non-finite values are filtered
Returns
-------
lnprob : float
Log posterior probability.
"""
# the comparison has to be done on theta and bounds. DO NOT inject theta
# values into Parameters, then compare Parameters values to the bounds.
# Parameters values are clipped to stay within bounds.
if np.any(theta > bounds[:, 1]) or np.any(theta < bounds[:, 0]):
return -np.inf
for name, val in zip(var_names, theta):
params[name].value = val
userkwargs = {}
if userkws is not None:
userkwargs = userkws
# update the constraints
params.update_constraints()
# now calculate the log-likelihood
out = userfcn(params, *userargs, **userkwargs)
out = _nan_policy(out, nan_policy=nan_policy, handle_inf=False)
lnprob = np.asarray(out).ravel()
if lnprob.size > 1:
# objective function returns a vector of residuals
if '__lnsigma' in params and not is_weighted:
# marginalise over a constant data uncertainty
__lnsigma = params['__lnsigma'].value
c = np.log(2 * np.pi) + 2 * __lnsigma
lnprob = -0.5 * np.sum((lnprob / np.exp(__lnsigma)) ** 2 + c)
else:
lnprob = -0.5 * (lnprob * lnprob).sum()
else:
# objective function returns a single value.
# use float_behaviour to figure out if the value is posterior or chi2
if float_behavior == 'posterior':
pass
elif float_behavior == 'chi2':
lnprob *= -0.5
else:
raise ValueError("float_behaviour must be either 'posterior' or"
" 'chi2' " + float_behavior)
return lnprob
def _make_random_gen(seed):
"""Turn seed into a numpy.random.RandomState instance.
If seed is None, return the RandomState singleton used by
numpy.random. If seed is an int, return a new RandomState instance
seeded with seed. If seed is already a RandomState instance, return
it. Otherwise raise ValueError.
"""
if seed is None or seed is np.random:
return np.random.mtrand._rand
if isinstance(seed, (numbers.Integral, np.integer)):
return np.random.RandomState(seed)
if isinstance(seed, np.random.RandomState):
return seed
raise ValueError('%r cannot be used to seed a numpy.random.RandomState'
' instance' % seed)
VALID_NAN_POLICIES = ('propagate', 'omit', 'raise')
def validate_nan_policy(policy):
"""Validate and rationalize `nan_policy`.
This function ensures backwards compatibility and as well as compatibility
with Pandas `missing` convention.
"""
if policy in VALID_NAN_POLICIES:
return policy
if policy is None:
policy = 'propagate'
policy = policy.lower()
if policy == 'drop':
policy = 'omit'
if policy == 'none':
policy = 'propagate'
if policy not in VALID_NAN_POLICIES:
raise ValueError("nan_policy must be 'propagate', 'omit', or 'raise'.")
return policy
def _nan_policy(arr, nan_policy='raise', handle_inf=True):
"""Specify behaviour when an array contains numpy.nan or numpy.inf.
Parameters
----------
arr : array_like
Input array to consider.
nan_policy : str, optional
One of:
'raise' - raise a `ValueError` if `arr` contains NaN (default)
'propagate' - propagate NaN
'omit' - filter NaN from input array
handle_inf : bool, optional
As well as NaN consider +/- inf.
Returns
-------
filtered_array : array_like
Note
----
This function is copied, then modified, from
scipy/stats/stats.py/_contains_nan
"""
nan_policy = validate_nan_policy(nan_policy)
if handle_inf:
handler_func = lambda x: ~np.isfinite(x)
else:
handler_func = isnull
if nan_policy == 'omit':
# mask locates any values to remove
mask = ~handler_func(arr)
if not np.all(mask): # there are some NaNs/infs/missing values
return arr[mask]
if nan_policy == 'raise':
try:
# Calling np.sum to avoid creating a huge array into memory
# e.g. np.isnan(a).any()
with np.errstate(invalid='ignore'):
contains_nan = handler_func(np.sum(arr))
except TypeError:
# If the check cannot be properly performed we fallback to omiting
# nan values and raising a warning. This can happen when attempting to
# sum things that are not numbers (e.g. as in the function `mode`).
contains_nan = False
warnings.warn("The input array could not be checked for NaNs. "
"NaNs will be ignored.", RuntimeWarning)
if contains_nan:
raise ValueError("The input contains nan values")
return arr
def minimize(fcn, params, method='leastsq', args=None, kws=None, iter_cb=None,
scale_covar=True, nan_policy='raise', reduce_fcn=None,
calc_covar=True, **fit_kws):
"""Perform a fit of a set of parameters by minimizing an objective (or
cost) function using one of the several available methods.
The minimize function takes an objective function to be minimized,
a dictionary (:class:`~lmfit.parameter.Parameters`) containing the model
parameters, and several optional arguments.
Parameters
----------
fcn : callable
Objective function to be minimized. When method is `leastsq` or
`least_squares`, the objective function should return an array
of residuals (difference between model and data) to be minimized
in a least-squares sense. With the scalar methods the objective
function can either return the residuals array or a single scalar
value. The function must have the signature:
`fcn(params, *args, **kws)`
params : :class:`~lmfit.parameter.Parameters`
Contains the Parameters for the model.
method : str, optional
Name of the fitting method to use. Valid values are:
- `'leastsq'`: Levenberg-Marquardt (default)
- `'least_squares'`: Least-Squares minimization, using Trust Region Reflective method
- `'differential_evolution'`: differential evolution
- `'brute'`: brute force method
- `'basinhopping'`: basinhopping
- `'ampgo'`: Adaptive Memory Programming for Global Optimization
- '`nelder`': Nelder-Mead
- `'lbfgsb'`: L-BFGS-B
- `'powell'`: Powell
- `'cg'`: Conjugate-Gradient
- `'newton'`: Newton-CG
- `'cobyla'`: Cobyla
- `'bfgs'`: BFGS
- `'tnc'`: Truncated Newton
- `'trust-ncg'`: Newton-CG trust-region
- `'trust-exact'`: nearly exact trust-region (SciPy >= 1.0)
- `'trust-krylov'`: Newton GLTR trust-region (SciPy >= 1.0)
- `'trust-constr'`: trust-region for constrained optimization (SciPy >= 1.1)
- `'dogleg'`: Dog-leg trust-region
- `'slsqp'`: Sequential Linear Squares Programming
- `'emcee'`: Maximum likelihood via Monte-Carlo Markov Chain
In most cases, these methods wrap and use the method of the same
name from `scipy.optimize`, or use `scipy.optimize.minimize` with
the same `method` argument. Thus '`leastsq`' will use
`scipy.optimize.leastsq`, while '`powell`' will use
`scipy.optimize.minimizer(..., method='powell')`
For more details on the fitting methods please refer to the
`SciPy docs <https://docs.scipy.org/doc/scipy/reference/optimize.html>`__.
args : tuple, optional
Positional arguments to pass to `fcn`.
kws : dict, optional
Keyword arguments to pass to `fcn`.
iter_cb : callable, optional
Function to be called at each fit iteration. This function should
have the signature `iter_cb(params, iter, resid, *args, **kws)`,
where `params` will have the current parameter values, `iter`
the iteration number, `resid` the current residual array, and `*args`
and `**kws` as passed to the objective function.
scale_covar : bool, optional
Whether to automatically scale the covariance matrix (default is True).
nan_policy : str, optional
Specifies action if `userfcn` (or a Jacobian) returns NaN
values. One of:
- 'raise' : a `ValueError` is raised
- 'propagate' : the values returned from `userfcn` are un-altered
- 'omit' : non-finite values are filtered
reduce_fcn : str or callable, optional
Function to convert a residual array to a scalar value for the scalar
minimizers. See notes in `Minimizer`.
calc_covar : bool, optional
Whether to calculate the covariance matrix (default is True) for
solvers other than `leastsq` and `least_squares`. Requires the
`numdifftools` package to be installed.
**fit_kws : dict, optional
Options to pass to the minimizer being used.
Returns
-------
:class:`MinimizerResult`
Object containing the optimized parameter and several
goodness-of-fit statistics.
.. versionchanged:: 0.9.0
Return value changed to :class:`MinimizerResult`.
Notes
-----
The objective function should return the value to be minimized. For the
Levenberg-Marquardt algorithm from leastsq(), this returned value must
be an array, with a length greater than or equal to the number of
fitting variables in the model. For the other methods, the return value
can either be a scalar or an array. If an array is returned, the sum of
squares of the array will be sent to the underlying fitting method,
effectively doing a least-squares optimization of the return values.
A common use for `args` and `kws` would be to pass in other
data needed to calculate the residual, including such things as the
data array, dependent variable, uncertainties in the data, and other
data structures for the model calculation.
On output, `params` will be unchanged. The best-fit values and, where
appropriate, estimated uncertainties and correlations, will all be
contained in the returned :class:`MinimizerResult`. See
:ref:`fit-results-label` for further details.
This function is simply a wrapper around :class:`Minimizer`
and is equivalent to::
fitter = Minimizer(fcn, params, fcn_args=args, fcn_kws=kws,
iter_cb=iter_cb, scale_covar=scale_covar,
nan_policy=nan_policy, reduce_fcn=reduce_fcn,
calc_covar=calc_covar, **fit_kws)
fitter.minimize(method=method)
"""
fitter = Minimizer(fcn, params, fcn_args=args, fcn_kws=kws,
iter_cb=iter_cb, scale_covar=scale_covar,
nan_policy=nan_policy, reduce_fcn=reduce_fcn,
calc_covar=calc_covar, **fit_kws)
return fitter.minimize(method=method)