def testModelSaveAndLoad(self):
# Keywords model uses the base class implementations of save/load methods.
self.modelDir = "poke_model"
model = ClassificationModelKeywords(modelDir=self.modelDir,
verbosity=0)
samples = {
0: (["Pickachu"], numpy.array([0, 2, 2])),
1: (["Eevee"], numpy.array([2])),
2: (["Charmander"], numpy.array([0, 1, 1])),
3: (["Abra"], numpy.array([1])),
4: (["Squirtle"], numpy.array([1, 0, 1]))
}
patterns = model.encodeSamples(samples)
for i in xrange(len(samples)):
model.trainModel(i)
output = [model.testModel(i) for i in xrange(len(patterns))]
model.saveModel()
loadedModel = ClassificationModel(verbosity=0).loadModel(self.modelDir)
loadedModelOutput = [
loadedModel.testModel(i) for i in xrange(len(patterns))
]
for mClasses, lClasses in zip(output, loadedModelOutput):
self.assertSequenceEqual(
mClasses.tolist(), lClasses.tolist(), "Output "
"classifcations from loaded model don't match original model's."
)
def testModelSaveAndLoad(self):
# Keywords model uses the base class implementations of save/load methods.
self.modelDir = "poke_model"
model = ClassificationModelKeywords(modelDir=self.modelDir, verbosity=0)
samples = {0: (["Pickachu"], numpy.array([0, 2, 2])),
1: (["Eevee"], numpy.array([2])),
2: (["Charmander"], numpy.array([0, 1, 1])),
3: (["Abra"], numpy.array([1])),
4: (["Squirtle"], numpy.array([1, 0, 1]))}
patterns = model.encodeSamples(samples)
for i in xrange(len(samples)):
model.trainModel(i)
output = [model.testModel(i) for i in xrange(len(patterns))]
model.saveModel()
loadedModel = ClassificationModel(verbosity=0).loadModel(self.modelDir)
loadedModelOutput = [loadedModel.testModel(i)
for i in xrange(len(patterns))]
for mClasses, lClasses in zip(output, loadedModelOutput):
self.assertSequenceEqual(mClasses.tolist(), lClasses.tolist(), "Output "
"classifcations from loaded model don't match original model's.")
def testNoWinningLabels(self):
"""Inferring 0/4 classes should return 0 winning labels."""
model = ClassificationModel()
inferenceResult = numpy.array([0, 0, 0, 0])
topLabels = model.getWinningLabels(inferenceResult)
self.assertFalse(topLabels)
def testNoWinningLabels(self):
"""Inferring 0/4 classes should return 0 winning labels."""
model = ClassificationModel()
inferenceResult = numpy.array([0, 0, 0, 0])
topLabels = model.getWinningLabels(inferenceResult)
self.assertFalse(topLabels)
def testCalculateAccuracyMultipleSamples(self):
"""
Tests testCalculateAccuracy() method of classification model base class for
three test samples.
"""
model = ClassificationModel()
actualLabels = [numpy.array([0]), numpy.array([0, 2]), numpy.array([0, 1, 2])]
predictedLabels = [numpy.array([0]), [None], numpy.array([1, 2, 0])]
classifications = [predictedLabels, actualLabels]
self.assertAlmostEqual(model.calculateAccuracy(classifications), float(2) / 3)
def testCalculateAccuracyMultipleSamples(self):
"""
Tests testCalculateAccuracy() method of classification model base class for
three test samples.
"""
model = ClassificationModel()
actualLabels = [
numpy.array([0]),
numpy.array([0, 2]),
numpy.array([0, 1, 2])
]
predictedLabels = [numpy.array([0]), [None], numpy.array([1, 2, 0])]
classifications = [predictedLabels, actualLabels]
self.assertAlmostEqual(model.calculateAccuracy(classifications),
float(2) / 3)
def testCalculateAccuracyMixedSamples(self):
"""
Tests testCalculateAccuracy() method of classification model base class for
test samples with mixed classifications.
"""
model = ClassificationModel()
actualLabels = [numpy.array([0, 1, 2])]
predictedLabels1 = [numpy.array([1, 2, 0])]
predictedLabels2 = [numpy.array([1])]
predictedLabels3 = [None]
classifications1 = [predictedLabels1, actualLabels]
classifications2 = [predictedLabels2, actualLabels]
classifications3 = [predictedLabels3, actualLabels]
self.assertAlmostEqual(model.calculateAccuracy(classifications1), 1.0)
self.assertAlmostEqual(model.calculateAccuracy(classifications2), float(1) / 3)
self.assertAlmostEqual(model.calculateAccuracy(classifications3), 0.0)
def testCalculateAccuracyMixedSamples(self):
"""
Tests testCalculateAccuracy() method of classification model base class for
test samples with mixed classifications.
"""
model = ClassificationModel()
actualLabels = [numpy.array([0, 1, 2])]
predictedLabels1 = [numpy.array([1, 2, 0])]
predictedLabels2 = [numpy.array([1])]
predictedLabels3 = [None]
classifications1 = [predictedLabels1, actualLabels]
classifications2 = [predictedLabels2, actualLabels]
classifications3 = [predictedLabels3, actualLabels]
self.assertAlmostEqual(model.calculateAccuracy(classifications1), 1.0)
self.assertAlmostEqual(model.calculateAccuracy(classifications2),
float(1) / 3)
self.assertAlmostEqual(model.calculateAccuracy(classifications3), 0.0)
def testWinningLabels(self):
"""
Tests whether classification base class returns multiple labels correctly.
"""
model = ClassificationModel()
inferenceResult = numpy.array([3, 1, 4, 0, 1, 0])
topLabels = model.getWinningLabels(inferenceResult, numLabels=1)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2])),
"Output should be label 2.")
topLabels = model.getWinningLabels(inferenceResult, numLabels=2)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0])),
"Output should be labels 2 and 0.")
# Test only nonzero labels are returned.
inferenceResult = numpy.array([3, 0, 4, 0, 0, 0])
topLabels = model.getWinningLabels(inferenceResult, numLabels=5)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0])),
"Output should be labels 2 and 0.")
def testWinningLabels(self):
"""
Tests whether classification base class returns multiple labels correctly.
"""
model = ClassificationModel()
inferenceResult = numpy.array([3, 1, 4, 0, 1, 0])
topLabels = model.getWinningLabels(inferenceResult, numLabels=1)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2])),
"Output should be label 2.")
topLabels = model.getWinningLabels(inferenceResult, numLabels=2)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0])),
"Output should be labels 2 and 0.")
# Test only nonzero labels are returned.
inferenceResult = numpy.array([3, 0, 4, 0, 0, 0])
topLabels = model.getWinningLabels(inferenceResult, numLabels=5)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0])),
"Output should be labels 2 and 0.")
def testWinningLabels(self):
"""
Tests whether classification base class returns multiple labels correctly.
"""
model = ClassificationModel()
inferenceResult = numpy.array([3, 1, 4, 0, 1, 0])
topLabels = model.getWinningLabels(inferenceResult, numLabels=1)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2])), "Output should be label 2.")
topLabels = model.getWinningLabels(inferenceResult, numLabels=2)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0])), "Output should be labels 2 and 0.")
# Test tie is broken by the later index; label 4 beats out label 1.
topLabels = model.getWinningLabels(inferenceResult, numLabels=3)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0, 4])), "Output should be labels 2, 0, and 1.")
# Test only nonzero labels are returned.
topLabels = model.getWinningLabels(inferenceResult, numLabels=5)
self.assertTrue(numpy.allclose(topLabels, numpy.array([2, 0, 4, 1])), "Output should be labels 2, 0, 1, and 4.")
