def test_predictions():
np.random.seed(222)
D = np.random.rand(100,10)
D[:40] += np.random.rand(40,10)**2
labels = [0] * 40 + [1] * 60
cmat,_,predictions = nfoldcrossvalidation(D, labels, classifier=fast_classifier(), return_predictions=1)
assert np.all((predictions == 0)|(predictions == 1))
assert cmat.trace() == np.sum(predictions == labels)
def test_nfoldcrossvalidation_simple():
from milksets import wine
features, labels = wine.load()
features = features[::2]
labels = labels[::2]
cmat,clabels = nfoldcrossvalidation(features, labels, classifier=fast_classifier())
assert cmat.shape == (3,3)
assert len(clabels) == 3
def test_stringlabels():
np.random.seed(222)
D = np.random.rand(100,10)
D[:40] += np.random.rand(40,10)**2
labelnames = ['one'] * 40 + ['two'] * 60
cmat,Lo = nfoldcrossvalidation(D, labelnames, classifier=fast_classifier())
assert Lo[0] in labelnames
assert Lo[1] in labelnames
assert Lo[0] != Lo[1] in labelnames
def test_nfoldcrossvalidation_simple():
from milksets import wine
features, labels = wine.load()
features = features[::2]
labels = labels[::2]
cmat, clabels = nfoldcrossvalidation(features,
labels,
classifier=fast_classifier())
assert cmat.shape == (3, 3)
assert len(clabels) == 3
def test_predictions():
np.random.seed(222)
D = np.random.rand(100, 10)
D[:40] += np.random.rand(40, 10)**2
labels = [0] * 40 + [1] * 60
cmat, _, predictions = nfoldcrossvalidation(D,
labels,
classifier=fast_classifier(),
return_predictions=1)
assert np.all((predictions == 0) | (predictions == 1))
assert cmat.trace() == np.sum(predictions == labels)
def test_stringlabels():
np.random.seed(222)
D = np.random.rand(100, 10)
D[:40] += np.random.rand(40, 10)**2
labelnames = ['one'] * 40 + ['two'] * 60
cmat, Lo = nfoldcrossvalidation(D,
labelnames,
classifier=fast_classifier())
assert Lo[0] in labelnames
assert Lo[1] in labelnames
assert Lo[0] != Lo[1] in labelnames
def test_tree():
mtree = milk.supervised.multi.multi_tree_learner(fast_classifier())
labels = [0,1,2,2,3,3,3,3]
features = np.random.random_sample((len(labels), 8))
model = mtree.train(features, labels)
counts = np.zeros(4)
for ell in labels:
counts[ell] += 1
g0,g1 = milk.supervised.multi.split(counts)
assert np.all(g0 == [3]) or np.all(g1 == [3])
def r(m):
if len(m) == 1: return int(m[0])
else: return sorted([r(m[1]), r(m[2])])
assert r(model.model) == [3,[2,[0,1]]]
def test_tree():
mtree = milk.supervised.multi.multi_tree_learner(fast_classifier())
labels = [0, 1, 2, 2, 3, 3, 3, 3]
features = np.random.random_sample((len(labels), 8))
model = mtree.train(features, labels)
counts = np.zeros(4)
for ell in labels:
counts[ell] += 1
g0, g1 = milk.supervised.multi.split(counts)
assert np.all(g0 == [3]) or np.all(g1 == [3])
def r(m):
if len(m) == 1: return int(m[0])
else: return sorted([r(m[1]), r(m[2])])
assert r(model.model) == [3, [2, [0, 1]]]
def test_classifier_no_set_options():
# Basically these should not raise an exception
milk.supervised.multi.one_against_rest_multi(fast_classifier())
milk.supervised.multi.one_against_rest(fast_classifier())
milk.supervised.multi.one_against_one(fast_classifier())
