def __init__(self, exog, model_gen, res='params', add_constant=True,
**kwargs):
"""
Parameters
----------
exog : array-like
Column ordered (observations in rows) design matrix.
model_gen : callable
Callable that returns a StatsModels model when called like
``model_gen(endog, exog)``.
res : {'params', 'tvalues', ...} or 1d array or 2d array or callable
Variable of interest that should be reported as feature-wise
measure. If a str, the corresponding attribute of the model fit result
class is returned (e.g. 'tvalues'). If a 1d-array, it is passed
to the fit result class' ``t_test()`` function as a t-contrast vector.
If a 2d-array, it is passed to the ``f_test()`` function as a
constrast matrix. If both latter cases a number of common test
statistics are return in the rows of the result dataset. A description
is available in the 'descr' sample attribute. Any other datatype
passed to this argument will be treated as a callable, the model
fit result is passed to it, and its return value(s) is aggregated
in the result dataset.
add_constant : bool
If True, a constant will be added to the design matrix that is
passed to ``exog``.
"""
FeaturewiseMeasure.__init__(self, **kwargs)
self._exog = exog
if add_constant:
self._exog = sm.add_constant(exog)
self._res = res
if isinstance(res, np.ndarray) or isinstance(res, (list, tuple)):
self._res = np.atleast_1d(res)
self._model_gen = model_gen
def __init__(self, exog, model_gen, res='params', add_constant=True,
**kwargs):
"""
Parameters
----------
exog : array-like
Column ordered (observations in rows) design matrix.
model_gen : callable
Callable that returns a StatsModels model when called like
``model_gen(endog, exog)``.
res : {'params', 'tvalues', ...} or 1d array or 2d array or callable
Variable of interest that should be reported as feature-wise
measure. If a str, the corresponding attribute of the model fit result
class is returned (e.g. 'tvalues'). If a 1d-array, it is passed
to the fit result class' ``t_test()`` function as a t-contrast vector.
If a 2d-array, it is passed to the ``f_test()`` function as a
contrast matrix. In both latter cases a number of common test
statistics are returned in the rows of the result dataset. A description
is available in the 'descr' sample attribute. Any other datatype
passed to this argument will be treated as a callable, the model
fit result is passed to it, and its return value(s) is aggregated
in the result dataset.
add_constant : bool, optional
If True, a constant will be added to the design matrix that is
passed to ``exog``.
"""
FeaturewiseMeasure.__init__(self, **kwargs)
self._exog = exog
if add_constant:
self._exog = sm.add_constant(exog)
self._res = res
if isinstance(res, np.ndarray) or isinstance(res, (list, tuple)):
self._res = np.atleast_1d(res)
self._model_gen = model_gen
def __init__(self, design, voi='pe', **kwargs):
"""
Parameters
----------
design : array (nsamples x nregressors)
GLM design matrix.
voi : {'pe', 'zstat'}
Variable of interest that should be reported as feature-wise
measure. 'beta' are the parameter estimates and 'zstat' returns
standardized parameter estimates.
"""
FeaturewiseMeasure.__init__(self, **kwargs)
# store the design matrix as a such (no copying if already array)
self._design = np.asmatrix(design)
# what should be computed ('variable of interest')
if not voi in ['pe', 'zstat']:
raise ValueError, \
"Unknown variable of interest '%s'" % str(voi)
self._voi = voi
# will store the precomputed Moore-Penrose pseudo-inverse of the
# design matrix (lazy calculation)
self._inv_design = None
# also store the inverse of the inner product for beta variance
# estimation
self._inv_ip = None
def __init__(self,
pvalue=False,
attr='targets',
corr_backend=None,
**kwargs):
"""Initialize
Parameters
----------
pvalue : bool
Either to report p-value of pearsons correlation coefficient
instead of pure correlation coefficient
attr : str
What attribut to correlate with
corr_backend: None or 'builtin' or 'scipy' (default: None)
Which function to use to compute correlations.
None means 'scipy' if pvalue else 'builtin'.
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__pvalue = int(pvalue)
self.__attr = attr
self.__corr_backend = corr_backend
def __init__(self, num_permutations=200, num_bootstraps=100, **kwargs):
raise NotImplemented, 'PLS was not yet implemented fully'
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
# save the args for the analysis
self.num_permutations = num_permutations
self.num_bootstraps = num_bootstraps
def __init__(self, num_permutations=200, num_bootstraps=100, **kwargs):
raise NotImplemented, 'PLS was not yet implemented fully'
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
# save the args for the analysis
self.num_permutations = num_permutations
self.num_bootstraps = num_bootstraps
def __init__(self, space='targets', **kwargs):
"""Initialize
Parameters
----------
space : str
What samples attribute to use as targets (labels).
"""
# init base classes first
FeaturewiseMeasure.__init__(self, space=space, **kwargs)
def __init__(self, space='targets', **kwargs):
"""Initialize
Parameters
----------
space : str
What samples attribute to use as targets (labels).
"""
# init base classes first
FeaturewiseMeasure.__init__(self, space=space, **kwargs)
def __init__(self, space='targets', **kwargs):
"""
Parameters
----------
space : str
What samples attribute to use as targets (labels).
"""
# set auto-train flag since we have nothing special to be done
# so by default auto train
kwargs['auto_train'] = kwargs.get('auto_train', True)
FeaturewiseMeasure.__init__(self, space=space, **kwargs)
def __init__(self, space='targets', **kwargs):
"""
Parameters
----------
space : str
What samples attribute to use as targets (labels).
"""
# set auto-train flag since we have nothing special to be done
# so by default auto train
kwargs['auto_train'] = kwargs.get('auto_train', True)
FeaturewiseMeasure.__init__(self, space=space, **kwargs)
def __init__(self, attr="targets", **kwargs):
"""Initialize
Parameters
----------
attr : str
Attribute to correlate across chunks.
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__attr = attr
def __init__(self, attr='targets', **kwargs):
"""Initialize
Parameters
----------
attr : str
Attribute to correlate across chunks.
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__attr = attr
def __init__(self, threshold=1.0e-2, kernel_width=1.0,
w_guess=None, **kwargs):
"""Constructor of the IRELIEF class.
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
# Threshold in W changes (stopping criterion for irelief).
self.threshold = threshold
self.w_guess = w_guess
self.w = None
self.kernel_width = kernel_width
def __init__(self, pvalue=False, attr='targets', **kwargs):
"""Initialize
Parameters
----------
pvalue : bool
Either to report p-value of pearsons correlation coefficient
instead of pure correlation coefficient
attr : str
What attribut to correlate with
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__pvalue = int(pvalue)
self.__attr = attr
def __init__(self, pvalue=False, attr='targets', **kwargs):
"""Initialize
Parameters
----------
pvalue : bool
Either to report p-value of pearsons correlation coefficient
instead of pure correlation coefficient
attr : str
What attribut to correlate with
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__pvalue = int(pvalue)
self.__attr = attr
def __init__(self, datameasure, noise=np.random.normal):
"""
Parameters
----------
datameasure : `Measure`
Used to quantify the effect of noise perturbation.
noise: Callable
Used to generate noise. The noise generator has to return an 1d array
of n values when called the `size=n` keyword argument. This is the
default interface of the random number generators in NumPy's
`random` module.
"""
# init base classes first
FeaturewiseMeasure.__init__(self)
self.__datameasure = datameasure
self.__noise = noise
def __init__(self, datameasure,
noise=np.random.normal):
"""
Parameters
----------
datameasure : `Measure`
Used to quantify the effect of noise perturbation.
noise: Callable
Used to generate noise. The noise generator has to return an 1d array
of n values when called the `size=n` keyword argument. This is the
default interface of the random number generators in NumPy's
`random` module.
"""
# init base classes first
FeaturewiseMeasure.__init__(self)
self.__datameasure = datameasure
self.__noise = noise
def __init__(self, pvalue=False, attr='targets',
corr_backend=None, **kwargs):
"""Initialize
Parameters
----------
pvalue : bool
Either to report p-value of pearsons correlation coefficient
instead of pure correlation coefficient
attr : str
What attribut to correlate with
corr_backend: None or 'builtin' or 'scipy' (default: None)
Which function to use to compute correlations.
None means 'scipy' if pvalue else 'builtin'.
"""
# init base classes first
FeaturewiseMeasure.__init__(self, **kwargs)
self.__pvalue = int(pvalue)
self.__attr = attr
self.__corr_backend = corr_backend
def __init__(self, mult=1, **kwargs):
FeaturewiseMeasure.__init__(self, **kwargs)
self.__mult = mult
def __init__(self, mult=1, **kwargs):
FeaturewiseMeasure.__init__(self, **kwargs)
self.__mult = mult
