Source Code for Module mvpa.tests.test_smlr

 1  # emacs: -*- mode: python; py-indent-offset: 4; indent-tabs-mode: nil -*- 
 2  # vi: set ft=python sts=4 ts=4 sw=4 et: 
 3  ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ## 
 4  # 
 5  #   See COPYING file distributed along with the PyMVPA package for the 
 6  #   copyright and license terms. 
 7  # 
 8  ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ## 
 9  """Unit tests for PyMVPA sparse multinomial logistic regression classifier""" 
10   
11  from mvpa.clfs.smlr import SMLR 
12  from tests_warehouse import * 
13  from mvpa.misc.data_generators import normalFeatureDataset 
14   
15   
16 -class SMLRTests(unittest.TestCase): 
17   
18 -    def testSMLR(self): 
19          data = datasets['dumb'] 
20   
21          clf = SMLR() 
22   
23          clf.train(data) 
24   
25          # prediction has to be perfect 
26          # 
27          # XXX yoh: whos said that?? ;-) 
28          # 
29          # There is always a tradeoff between learning and 
30          # generalization errors so...  but in this case the problem is 
31          # more interesting: absent bias disallows to learn data you 
32          # have here -- there is no solution which would pass through 
33          # (0,0) 
34          predictions = clf.predict(data.samples) 
35          self.failUnless((predictions == data.labels).all()) 
36   
37   
38 -    def testSMLRState(self): 
39          data = datasets['dumb'] 
40   
41          clf = SMLR() 
42   
43          clf.train(data) 
44   
45          clf.states.enable('values') 
46          clf.states.enable('predictions') 
47   
48          p = N.asarray(clf.predict(data.samples)) 
49   
50          self.failUnless((p == clf.predictions).all()) 
51          self.failUnless(N.array(clf.values).shape[0] == N.array(p).shape[0]) 
52   
53   
54 -    def testSMLRSensitivities(self): 
55          data = normalFeatureDataset(perlabel=10, nlabels=2, nfeatures=4) 
56   
57          # use SMLR on binary problem, but not fitting all weights 
58          clf = SMLR(fit_all_weights=False) 
59          clf.train(data) 
60   
61          # now ask for the sensitivities WITHOUT having to pass the dataset 
62          # again 
63          sens = clf.getSensitivityAnalyzer(force_training=False)() 
64   
65          self.failUnless(sens.shape == (data.nfeatures,)) 
66   
67   
68 -def suite(): 
69      return unittest.makeSuite(SMLRTests) 
70   
71   
72  if __name__ == '__main__': 
73      import runner 
74