/usr/share/pyshared/mlpy/_srda.py is in python-mlpy 2.2.0~dfsg1-2.1.
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 | ## Spectral Regression Discriminant Analysis.
## This is an implementation of Spectral Regression Discriminant Analysis described in:
## 'SRDA: An Efficient Algorithm for Large ScaleDiscriminant Analysis' Deng Cai,
## Xiaofei He, Jiawei Han. 2008.
## This code is written by Roberto Visintainer, <visintainer@fbk.eu> and Davide Albanese, <albanese@fbk.eu>.
## (C) 2008 Fondazione Bruno Kessler - Via Santa Croce 77, 38100 Trento, ITALY.
## This program is free software: you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation, either version 3 of the License, or
## (at your option) any later version.
## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
## GNU General Public License for more details.
## You should have received a copy of the GNU General Public License
## along with this program. If not, see <http://www.gnu.org/licenses/>.
__all__ = ['Srda']
from numpy import *
from numpy.linalg import inv
class Srda:
"""Spectral Regression Discriminant Analysis (SRDA).
Example:
>>> import numpy as np
>>> import mlpy
>>> xtr = np.array([[1.0, 2.0, 3.1, 1.0], # first sample
... [1.0, 2.0, 3.0, 2.0], # second sample
... [1.0, 2.0, 3.1, 1.0]]) # third sample
>>> ytr = np.array([1, -1, 1]) # classes
>>> mysrda = mlpy.Srda() # initialize srda class
>>> mysrda.compute(xtr, ytr) # compute srda
1
>>> mysrda.predict(xtr) # predict srda model on training data
array([ 1, -1, 1])
>>> xts = np.array([4.0, 5.0, 6.0, 7.0]) # test point
>>> mysrda.predict(xts) # predict srda model on test point
-1
>>> mysrda.realpred # real-valued prediction
-6.8283034257748758
>>> mysrda.weights(xtr, ytr) # compute weights on training data
array([ 0.10766721, 0.21533442, 0.51386623, 1.69331158])
"""
def __init__ (self, alpha = 1.0):
"""Initialize the Srda class.
:Parameters:
alpha : float(>=0.0)
regularization parameter
"""
if alpha < 0.0:
raise ValueError("alpha (regularization parameter) must be >= 0.0")
self.__alpha = alpha
self.__classes = None
self.__a = None
self.__th = 0.0
self.__computed = False
self.realpred = None
def compute (self, x, y):
"""
Compute Srda model.
Initialize array of alphas a.
:Parameters:
x : 2d ndarray float (samples x feats)
training data
y : 1d ndarray integer (-1 or 1)
classes
:Returns:
1
:Raises:
LinAlgError
if x is singular matrix in __PenRegrModel
"""
# See eq 19 and 24
self.__classes = unique(y)
if self.__classes.shape[0] != 2:
raise ValueError("SRDA works only for two-classes problems")
cl0 = where(y == self.__classes[0])[0]
cl1 = where(y == self.__classes[1])[0]
ncl0 = cl0.shape[0]
ncl1 = cl1.shape[0]
y0 = x.shape[0] / float(ncl0)
y1 = -x.shape[0] / float(ncl1)
ym = append(ones(ncl0) * y0, ones(ncl1) * y1, axis = 1)
newpos = r_[cl0, cl1]
xi = x[newpos]
xiT = xi.transpose()
xXI = inv(dot(xi, xiT) + 1.0 + (self.__alpha * identity(x.shape[0])))
c = dot(xXI, ym)
self.__sumC = sum(c)
self.__a = dot(xiT, c)
##### Threshold tuning ######
ncomptrue = empty(x.shape[0], dtype = int)
ths = empty(x.shape[0])
ytmp = empty_like(y)
self.__computed = True
self.predict(x)
rpsorted = sort(self.__rp_noTh)
for t in range(ths.shape[0] - 1):
ths[t] = (rpsorted[t] + rpsorted[t + 1]) * 0.5
ytmp[self.__rp_noTh <= ths[t]] = self.__classes[0]
ytmp[self.__rp_noTh > ths[t]] = self.__classes[1]
comp = (y == ytmp)
ncomptrue[t] = sum(comp)
# Try th = 0.0
ths[-1] = 0.0
ytmp[self.__rp_noTh <= ths[-1]] = self.__classes[0]
ytmp[self.__rp_noTh > ths[-1]] = self.__classes[1]
comp = (y == ytmp)
ncomptrue[-1] = sum(comp)
self.__th = ths[argmax(ncomptrue)]
#############################
return 1
def weights (self, x, y):
"""Return feature weights.
:Parameters:
x : 2d ndarray float (samples x feats)
training data
y : 1d ndarray integer (-1 or 1)
classes
:Returns:
fw : 1d ndarray float
feature weights
"""
self.compute(x, y)
return abs(self.__a)
def predict (self, p):
"""Predict Srda model on test point(s).
:Parameters:
p : 1d or 2d ndarray float (sample(s) x feats)
test sample(s)
:Returns:
cl : integer or 1d numpy array integer
class(es) predicted
:Attributes:
self.realpred : float or 1d numpy array float
real valued prediction
"""
if self.__computed == False:
raise StandardError("No SRDA model computed")
if p.ndim == 2:
pred = empty((p.shape[0]), int)
self.__rp_noTh = -dot(self.__a, p.transpose()) - self.__sumC
self.realpred = self.__rp_noTh - self.__th
pred[self.realpred <= 0.0] = self.__classes[0]
pred[self.realpred > 0.0] = self.__classes[1]
return pred
elif p.ndim == 1:
self.__rp_noTh = -dot(p, self.__a) - self.__sumC
self.realpred = self.__rp_noTh - self.__th
if self.realpred <= 0.0: pred = self.__classes[0]
elif self.realpred > 0.0: pred = self.__classes[1]
return pred
|