def test_simple_multiclass_classification(self):
"""
simple LibSvmClassifier test - 3-class
Test libSVM classification functionality using random constructed
data, training the y=0.33 and y=.66 split
"""
DIM = 2
N = 1000
P1_LABEL = 'p1'
P2_LABEL = 'p2'
P3_LABEL = 'p3'
p = multiprocessing.pool.ThreadPool()
d_factory = DescriptorElementFactory(DescriptorMemoryElement, {})
di = 0
def make_element(iv):
i, v = iv
elem = d_factory.new_descriptor('test', i)
elem.set_vector(v)
return elem
# Constructing artificial descriptors
x = numpy.random.rand(N, DIM)
x_p1 = x[x[:, 1] <= 0.30]
x_p2 = x[(x[:, 1] >= 0.36) & (x[:, 1] <= 0.63)]
x_p3 = x[x[:, 1] >= 0.69]
d_p1 = p.map(make_element, enumerate(x_p1, di))
di += len(d_p1)
d_p2 = p.map(make_element, enumerate(x_p2, di))
di += len(d_p2)
d_p3 = p.map(make_element, enumerate(x_p3, di))
di += len(d_p3)
# Create/Train test classifier
classifier = LibSvmClassifier(
train_params={
'-t': 0, # linear kernel
'-b': 1, # enable probability estimates
'-c': 2, # SVM-C parameter C
'-q': '' # quite mode
},
normalize=None, # DO NOT normalize descriptors
)
classifier.train({P1_LABEL: d_p1, P2_LABEL: d_p2, P3_LABEL: d_p3})
# Test classifier
x = numpy.random.rand(N, DIM)
x_p1 = x[x[:, 1] <= 0.30]
x_p2 = x[(x[:, 1] >= 0.36) & (x[:, 1] <= 0.63)]
x_p3 = x[x[:, 1] >= 0.69]
# Test that examples expected to classify to certain classes are.
c_map_p1 = list(classifier._classify_arrays(x_p1))
for v, c_map in zip(x_p1, c_map_p1):
assert c_map[P1_LABEL] > max(c_map[P2_LABEL], c_map[P3_LABEL]), \
"Incorrect {} label: {} :: {}".format(P1_LABEL, v, c_map)
c_map_p2 = list(classifier._classify_arrays(x_p2))
for v, c_map in zip(x_p2, c_map_p2):
assert c_map[P2_LABEL] > max(c_map[P1_LABEL], c_map[P3_LABEL]), \
"Incorrect {} label: {} :: {}".format(P2_LABEL, v, c_map)
c_map_p3 = list(classifier._classify_arrays(x_p3))
for v, c_map in zip(x_p3, c_map_p3):
assert c_map[P3_LABEL] > max(c_map[P1_LABEL], c_map[P2_LABEL]), \
"Incorrect {} label: {} :: {}".format(P3_LABEL, v, c_map)
# Closing resources
p.close()
p.join()
def test_no_save_model_pickle(self):
# Test model preservation across pickling even without model cache
# file paths set.
classifier = LibSvmClassifier(
train_params={
'-t': 0, # linear kernel
'-b': 1, # enable probability estimates
'-c': 2, # SVM-C parameter C
'-q': '', # quite mode
},
normalize=None, # DO NOT normalize descriptors
)
self.assertTrue(classifier.svm_model is None)
# Empty model should not trigger __LOCAL__ content in pickle
self.assertNotIn('__LOCAL__', classifier.__getstate__())
_ = cPickle.loads(cPickle.dumps(classifier))
# train arbitrary model (same as ``test_simple_classification``)
DIM = 2
N = 1000
POS_LABEL = 'positive'
NEG_LABEL = 'negative'
d_factory = DescriptorElementFactory(DescriptorMemoryElement, {})
c_factory = ClassificationElementFactory(
MemoryClassificationElement, {}
)
def make_element(iv):
i, v = iv
d = d_factory.new_descriptor('test', i)
d.set_vector(v)
return d
# Constructing artificial descriptors
x = numpy.random.rand(N, DIM)
x_pos = x[x[:, 1] <= 0.45]
x_neg = x[x[:, 1] >= 0.55]
p = multiprocessing.pool.ThreadPool()
d_pos = p.map(make_element, enumerate(x_pos))
d_neg = p.map(make_element, enumerate(x_neg, start=N//2))
p.close()
p.join()
# Training
classifier.train({POS_LABEL: d_pos, NEG_LABEL: d_neg})
# Test original classifier
# - Using classification method implemented by the subclass directly
# in order to test simplest scope possible.
t_v = numpy.random.rand(DIM)
c_expected = list(classifier._classify_arrays([t_v]))[0]
# Should see __LOCAL__ content in pickle state now
p_state = classifier.__getstate__()
self.assertIn('__LOCAL__', p_state)
self.assertIn('__LOCAL_LABELS__', p_state)
self.assertIn('__LOCAL_MODEL__', p_state)
self.assertTrue(len(p_state['__LOCAL_LABELS__']) > 0)
self.assertTrue(len(p_state['__LOCAL_MODEL__']) > 0)
# Restored classifier should classify the same test descriptor the
# same
#: :type: LibSvmClassifier
classifier2 = cPickle.loads(cPickle.dumps(classifier))
c_post_pickle = list(classifier2._classify_arrays([t_v]))[0]
# There may be floating point error, so extract actual confidence
# values and check post round
c_pp_positive = c_post_pickle[POS_LABEL]
c_pp_negative = c_post_pickle[NEG_LABEL]
c_e_positive = c_expected[POS_LABEL]
c_e_negative = c_expected[NEG_LABEL]
self.assertAlmostEqual(c_e_positive, c_pp_positive, 5)
self.assertAlmostEqual(c_e_negative, c_pp_negative, 5)
def test_simple_classification(self):
"""
simple LibSvmClassifier test - 2-class
Test libSVM classification functionality using random constructed
data, training the y=0.5 split
"""
DIM = 2
N = 1000
POS_LABEL = 'positive'
NEG_LABEL = 'negative'
p = multiprocessing.pool.ThreadPool()
d_factory = DescriptorElementFactory(DescriptorMemoryElement, {})
def make_element(iv):
i, v = iv
elem = d_factory.new_descriptor('test', i)
elem.set_vector(v)
return elem
# Constructing artificial descriptors
x = numpy.random.rand(N, DIM)
x_pos = x[x[:, 1] <= 0.45]
x_neg = x[x[:, 1] >= 0.55]
d_pos = p.map(make_element, enumerate(x_pos))
d_neg = p.map(make_element, enumerate(x_neg, start=N//2))
# Create/Train test classifier
classifier = LibSvmClassifier(
train_params={
'-t': 0, # linear kernel
'-b': 1, # enable probability estimates
'-c': 2, # SVM-C parameter C
'-q': '', # quite mode
},
normalize=None, # DO NOT normalize descriptors
)
classifier.train({POS_LABEL: d_pos, NEG_LABEL: d_neg})
# Test classifier
x = numpy.random.rand(N, DIM)
x_pos = x[x[:, 1] <= 0.45]
x_neg = x[x[:, 1] >= 0.55]
# Test that examples expected to classify to the positive class are,
# and same for those expected to be in the negative class.
c_map_pos = list(classifier._classify_arrays(x_pos))
for v, c_map in zip(x_pos, c_map_pos):
assert c_map[POS_LABEL] > c_map[NEG_LABEL], \
"Found False positive: {} :: {}" \
.format(v, c_map)
c_map_neg = list(classifier._classify_arrays(x_neg))
for v, c_map in zip(x_neg, c_map_neg):
assert c_map[NEG_LABEL] > c_map[POS_LABEL], \
"Found False negative: {} :: {}" \
.format(v, c_map)
# Closing resources
p.close()
p.join()
