Package mvpa :: Package tests :: Module test_clf :: Class ClassifiersTests
[hide private]
[frames] | no frames]

Class ClassifiersTests

source code


Instance Methods [hide private]
 
setUp(self) source code
 
testDummy(self) source code
 
testBoosted(self) source code
 
testBoostedStatePropagation(self) source code
 
testBinaryDecorator(self) source code
 
testClassifierGeneralization(self, clf)
Simple test if classifiers can generalize ok on simple data
source code
 
testSummary(self, clf)
Basic testing of the clf summary
source code
 
testDegenerateUsage(self, clf)
Test how clf handles degenerate cases
source code
 
test_single_class(self, clf)
Test if binary and multiclass can handle single class training/testing
source code
 
testSplitClassifier(self) source code
 
testSplitClassifierExtended(self, clf_) source code
 
testHarvesting(self)
Basic testing of harvesting based on SplitClassifier
source code
 
testMappedClassifier(self) source code
 
testFeatureSelectionClassifier(self) source code
 
testFeatureSelectionClassifierWithRegression(self) source code
 
testTreeClassifier(self)
Basic tests for TreeClassifier
source code
 
testValues(self, clf) source code
 
testMulticlassClassifier(self, clf) source code
 
testSVMs(self, clf) source code
 
testRetrainables(self, clf) source code
 
testGenericTests(self)
Test all classifiers for conformant behavior
source code
 
testCorrectDimensionsOrder(self, clf)
To check if known/present Classifiers are working properly with samples being first dimension. Started to worry about possible problems while looking at sg where samples are 2nd dimension
source code
Method Details [hide private]

testClassifierGeneralization(self, clf)

source code 
Simple test if classifiers can generalize ok on simple data
Decorators:
  • @sweepargs(clf= clfswh ['binary'])

testSummary(self, clf)

source code 
Basic testing of the clf summary
Decorators:
  • @sweepargs(clf= clfswh [:]+ regrswh [:])

testDegenerateUsage(self, clf)

source code 
Test how clf handles degenerate cases
Decorators:
  • @sweepargs(clf= clfswh [:]+ regrswh [:])

test_single_class(self, clf)

source code 
Test if binary and multiclass can handle single class training/testing
Decorators:
  • @sweepargs(clf= clfswh ['!sg', '!plr', '!meta'])

testSplitClassifierExtended(self, clf_)

source code 
Decorators:
  • @sweepargs(clf_= clfswh ['binary', '!meta'])

testValues(self, clf)

source code 
Decorators:
  • @sweepargs(clf= clfswh [:])

testMulticlassClassifier(self, clf)

source code 
Decorators:
  • @sweepargs(clf= clfswh ['linear', 'svm', 'libsvm', '!meta'])

testSVMs(self, clf)

source code 
Decorators:
  • @sweepargs(clf= clfswh ['svm', '!meta'])

testRetrainables(self, clf)

source code 
Decorators:
  • @sweepargs(clf= clfswh ['retrainable'])

testCorrectDimensionsOrder(self, clf)

source code 
To check if known/present Classifiers are working properly with samples being first dimension. Started to worry about possible problems while looking at sg where samples are 2nd dimension
Decorators:
  • @sweepargs(clf= clfswh ['!smlr', '!knn', '!gnb', '!lars', '!meta', '!ridge'])