/usr/share/pyshared/sklearn/qda.py is in python-sklearn 0.14.1-2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 | """
Quadratic Discriminant Analysis
"""
# Author: Matthieu Perrot <matthieu.perrot@gmail.com>
#
# License: BSD 3 clause
import warnings
import numpy as np
from .base import BaseEstimator, ClassifierMixin
from .externals.six.moves import xrange
from .utils.fixes import unique
from .utils import check_arrays, array2d, column_or_1d
__all__ = ['QDA']
class QDA(BaseEstimator, ClassifierMixin):
"""
Quadratic Discriminant Analysis (QDA)
A classifier with a quadratic decision boundary, generated
by fitting class conditional densities to the data
and using Bayes' rule.
The model fits a Gaussian density to each class.
Parameters
----------
priors : array, optional, shape = [n_classes]
Priors on classes
reg_param : float, optional
Regularizes the covariance estimate as
``(1-reg_param)*Sigma + reg_param*np.eye(n_features)``
Attributes
----------
`covariances_` : list of array-like, shape = [n_features, n_features]
Covariance matrices of each class.
`means_` : array-like, shape = [n_classes, n_features]
Class means.
`priors_` : array-like, shape = [n_classes]
Class priors (sum to 1).
`rotations_` : list of arrays
For each class an array of shape [n_samples, n_samples], the
rotation of the Gaussian distribution, i.e. its principal axis.
`scalings_` : array-like, shape = [n_classes, n_features]
Contains the scaling of the Gaussian
distributions along the principal axes for each
class, i.e. the variance in the rotated coordinate system.
Examples
--------
>>> from sklearn.qda import QDA
>>> import numpy as np
>>> X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])
>>> y = np.array([1, 1, 1, 2, 2, 2])
>>> clf = QDA()
>>> clf.fit(X, y)
QDA(priors=None, reg_param=0.0)
>>> print(clf.predict([[-0.8, -1]]))
[1]
See also
--------
sklearn.lda.LDA: Linear discriminant analysis
"""
def __init__(self, priors=None, reg_param=0.):
self.priors = np.asarray(priors) if priors is not None else None
self.reg_param = reg_param
def fit(self, X, y, store_covariances=False, tol=1.0e-4):
"""
Fit the QDA model according to the given training data and parameters.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Training vector, where n_samples in the number of samples and
n_features is the number of features.
y : array, shape = [n_samples]
Target values (integers)
store_covariances : boolean
If True the covariance matrices are computed and stored in the
`self.covariances_` attribute.
"""
X, y = check_arrays(X, y)
y = column_or_1d(y, warn=True)
self.classes_, y = unique(y, return_inverse=True)
n_samples, n_features = X.shape
n_classes = len(self.classes_)
if n_classes < 2:
raise ValueError('y has less than 2 classes')
if self.priors is None:
self.priors_ = np.bincount(y) / float(n_samples)
else:
self.priors_ = self.priors
cov = None
if store_covariances:
cov = []
means = []
scalings = []
rotations = []
for ind in xrange(n_classes):
Xg = X[y == ind, :]
meang = Xg.mean(0)
means.append(meang)
Xgc = Xg - meang
# Xgc = U * S * V.T
U, S, Vt = np.linalg.svd(Xgc, full_matrices=False)
rank = np.sum(S > tol)
if rank < n_features:
warnings.warn("Variables are collinear")
S2 = (S ** 2) / (len(Xg) - 1)
S2 = ((1 - self.reg_param) * S2) + self.reg_param
if store_covariances:
# cov = V * (S^2 / (n-1)) * V.T
cov.append(np.dot(S2 * Vt.T, Vt))
scalings.append(S2)
rotations.append(Vt.T)
if store_covariances:
self.covariances_ = cov
self.means_ = np.asarray(means)
self.scalings_ = np.asarray(scalings)
self.rotations_ = rotations
return self
@property
def scalings(self): # pragma: no cover
warnings.warn("QDA.scalings is deprecated and will be removed in 0.15."
" Use QDA.scalings_ instead.", DeprecationWarning,
stacklevel=2)
return self.scalings_
@property
def rotations(self): # pragma: no cover
warnings.warn("QDA.rotations is deprecated and will be removed in "
"0.15. Use QDA.rotations_ instead.", DeprecationWarning,
stacklevel=2)
return self.rotations_
def _decision_function(self, X):
X = array2d(X)
norm2 = []
for i in range(len(self.classes_)):
R = self.rotations_[i]
S = self.scalings_[i]
Xm = X - self.means_[i]
X2 = np.dot(Xm, R * (S ** (-0.5)))
norm2.append(np.sum(X2 ** 2, 1))
norm2 = np.array(norm2).T # shape = [len(X), n_classes]
return (-0.5 * (norm2 + np.sum(np.log(self.scalings_), 1))
+ np.log(self.priors_))
def decision_function(self, X):
"""Apply decision function to an array of samples.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Array of samples (test vectors).
Returns
-------
C : array, shape = [n_samples, n_classes] or [n_samples,]
Decision function values related to each class, per sample.
In the two-class case, the shape is [n_samples,], giving the
log likelihood ratio of the positive class.
"""
dec_func = self._decision_function(X)
# handle special case of two classes
if len(self.classes_) == 2:
return dec_func[:, 1] - dec_func[:, 0]
return dec_func
def predict(self, X):
"""Perform classification on an array of test vectors X.
The predicted class C for each sample in X is returned.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
C : array, shape = [n_samples]
"""
d = self._decision_function(X)
y_pred = self.classes_.take(d.argmax(1))
return y_pred
def predict_proba(self, X):
"""Return posterior probabilities of classification.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Array of samples/test vectors.
Returns
-------
C : array, shape = [n_samples, n_classes]
Posterior probabilities of classification per class.
"""
values = self._decision_function(X)
# compute the likelihood of the underlying gaussian models
# up to a multiplicative constant.
likelihood = np.exp(values - values.max(axis=1)[:, np.newaxis])
# compute posterior probabilities
return likelihood / likelihood.sum(axis=1)[:, np.newaxis]
def predict_log_proba(self, X):
"""Return posterior probabilities of classification.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Array of samples/test vectors.
Returns
-------
C : array, shape = [n_samples, n_classes]
Posterior log-probabilities of classification per class.
"""
# XXX : can do better to avoid precision overflows
probas_ = self.predict_proba(X)
return np.log(probas_)
|