Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

PR: inverse_transform is implemented for scikit-learn utility #12

Merged
merged 2 commits into from Aug 22, 2020
Merged

PR: inverse_transform is implemented for scikit-learn utility #12

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d5ec27b
feat: inverse_transform is implemented
sshojiro May 1, 2020
3d1086e
refactor: sklearn eGTM now can be incorporated into `sklearn.pipeline…
sshojiro May 2, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
69 changes: 43 additions & 26 deletions ugtm/ugtm_sklearn.py
View file
Open in desktop
Expand Up @@ -41,12 +41,18 @@ class eGTM(BaseEstimator, TransformerMixin):
Number of iterations for EM algorithm.
verbose : bool, optional (default = False)
Verbose mode (outputs loglikelihood values during EM algorithm).
model : {'means', 'modes', 'responsibilities','complete'}, optional
GTM data representations:
'means' for mean data positions,
'modes' for positions with max. responsibilities,
'responsibilities' for probability distribution on the map,
'complete' for a complete instance of :class:`~ugtm.ugtm_classes.OptimizedGTM`

"""

def __init__(self, k=16, m=4, s=0.3, regul=0.1,
random_state=1234,
niter=200, verbose=False):
niter=200, verbose=False, model="means"):
"""Constructor for eGTM class.

Parameters
Expand Down Expand Up @@ -75,16 +81,26 @@ def __init__(self, k=16, m=4, s=0.3, regul=0.1,
Number of iterations for EM algorithm.
verbose : bool, optional (default = False)
Verbose mode (outputs loglikelihood values during EM algorithm).
model : {'means', 'modes', 'responsibilities','complete'}, optional
GTM data representations:
'means' for mean data positions,
'modes' for positions with max. responsibilities,
'responsibilities' for probability distribution on the map,
'complete' for a complete instance of :class:`~ugtm.ugtm_classes.OptimizedGTM`

"""
assert model in ('means', 'modes', 'responsibilities','complete'),\
"model must be either of 'means', 'modes', 'responsibilities', or 'complete'"
self.k = k
self.m = m
self.s = s
self.regul = regul
self.random_state = random_state
self.niter = niter
self.verbose = verbose
self.model = model

def fit(self, X):
def fit(self, X, y=None):
"""Fits GTM to X using :class:`~ugtm.ugtm_classes.OptimizedGTM`.

Parameters
Expand All @@ -106,27 +122,21 @@ def fit(self, X):

return self

def transform(self, X, model="means"):
def transform(self, X):
"""Projects new data X onto GTM using :func:`~ugtm.ugtm_gtm.projection`.

Parameters
==========

X : 2D array
Data matrix.
model : {'means', 'modes', 'responsibilities','complete'}, optional
GTM data representations:
'means' for mean data positions,
'modes' for positions with max. responsibilities,
'responsibilities' for probability distribution on the map,
'complete' for a complete instance of :class:`~ugtm.ugtm_classes.OptimizedGTM`

Returns
=======
if model="means", array of shape (n_instances, 2),
if model="modes", array of shape (n_instances, 2),
if model="responsibilities", array of shape (n_instances, n_nodes),
if model="complete", instance of class :class:`~ugtm.ugtm_classes.OptimizedGTM`
if self.model="means", array of shape (n_instances, 2),
if self.model="modes", array of shape (n_instances, 2),
if self.model="responsibilities", array of shape (n_instances, n_nodes),
if self.model="complete", instance of class :class:`~ugtm.ugtm_classes.OptimizedGTM`
"""

# Check fit
Expand All @@ -145,29 +155,23 @@ def transform(self, X, model="means"):
dic["modes"] = self.projected.matModes
dic["responsibilities"] = self.projected.matR

return dic[model]
return dic[self.model]

def fit_transform(self, X, model="means"):
def fit_transform(self, X, y=None):
"""Fits and transforms X using GTM.

Parameters
==========

X : 2D array
Data matrix.
model : {'means', 'modes', 'responsibilities','complete'}, optional
GTM data representations:
'means' for mean data positions,
'modes' for positions with max. responsibilities,
'responsibilities' for probability distribution on the map,
'complete' for a complete instance of :class:`~ugtm.ugtm_classes.OptimizedGTM`

Returns
=======
if model="means", array of shape (n_instances, 2),
if model="modes", array of shape (n_instances, 2),
if model="responsibilities", array of shape (n_instances, n_nodes),
if model="complete", instance of class :class:`~ugtm.ugtm_classes.OptimizedGTM`
if self.model="means", array of shape (n_instances, 2),
if self.model="modes", array of shape (n_instances, 2),
if self.model="responsibilities", array of shape (n_instances, n_nodes),
if self.model="complete", instance of class :class:`~ugtm.ugtm_classes.OptimizedGTM`
"""

X = check_array(X)
Expand Down Expand Up @@ -196,8 +200,21 @@ def fit_transform(self, X, model="means"):
dic["means"] = self.projected.matMeans
dic["modes"] = self.projected.matModes
dic["responsibilities"] = self.projected.matR
return dic[self.model]

def inverse_transform(self, matR):
"""Inverse transformation of responsibility onto the original data space

return dic[model]
Parameters
==========
matR : array of shape (n_samples, n_nodes)

Returns
=======
matY : array of shape (n_samples, n_dimensions)
"""
weightedPhi = np.dot(matR, self.initialModel.matPhiMPlusOne)
return np.dot(weightedPhi, self.optimizedModel.matW.T)


class eGTC(BaseEstimator, ClassifierMixin):
Expand Down