def test_SimpleTree_classification_pickle(self):
lrn = SimpleTreeCls()
clf = lrn(self.data_cls)
p = clf(self.data_cls, clf.Probs)
clf_ = pickle.loads(pickle.dumps(clf))
p_ = clf_(self.data_cls, clf.Probs)
np.testing.assert_almost_equal(p, p_)
def test_SimpleTree_single_instance(self):
data = Orange.data.Table('iris')
lrn = SimpleTreeCls()
clf = lrn(data)
for ins in data[::20]:
clf(ins)
val, prob = clf(ins, clf.ValueProbs)
self.assertEqual(sum(prob[0]), 1)
def test_SimpleTree_classification(self):
lrn = SimpleTreeCls()
clf = lrn(self.data_cls)
p = clf(self.data_cls, clf.Probs)
self.assertEqual(p.shape, (self.N, self.cls_vals))
self.assertAlmostEqual(p.min(), 0)
self.assertAlmostEqual(p.max(), 1)
np.testing.assert_almost_equal(p.sum(axis=1), np.ones(self.N))
def test_SimpleTree_classification_tree(self):
lrn = SimpleTreeCls(min_instances=6, max_majority=0.7)
clf = lrn(self.data_cls)
self.assertEqual(
clf.dumps_tree(clf.node),
'{ 1 4 -1.17364 { 1 5 0.37564 { 2 0.00 0.00 0.56 } '
'{ 2 0.00 3.00 1.14 } } { 1 4 -0.41863 { 1 5 0.14592 '
'{ 2 3.54 0.54 0.70 } { 2 2.46 0.46 2.47 } } { 1 4 0.24404 '
'{ 1 4 0.00654 { 1 3 -0.15750 { 2 1.00 0.00 0.45 } '
'{ 2 1.00 3.00 0.48 } } { 2 1.00 5.00 0.70 } } { 1 5 0.32635 '
'{ 2 0.52 2.52 4.21 } { 2 2.48 3.48 1.30 } } } } }')
def test_SimpleTree_to_string_classification(self):
domain = Domain([
DiscreteVariable(name='d1', values='ef'),
ContinuousVariable(name='c1')
], DiscreteVariable(name='cls', values='abc'))
data = Table(domain, [['e', 1, 'a'], ['e', 1, 'b'], ['e', 2, 'b'],
['f', 2, "c"], ["e", 3, "a"], ['f', 3, "c"]])
lrn = SimpleTreeCls(min_instances=1)
clf = lrn(data)
clf_str = clf.to_string()
res = '\nd1 ([2.0, 2.0, 2.0])\n: e\n c1 ([2.0, 2.0, 0.0])\n : <=2.5\n c1 ([1.0, 2.0, 0.0])\n : <=1.5 --> a ([1.0, 1.0, 0.0])\n : >1.5 --> b ([0.0, 1.0, 0.0])\n : >2.5 --> a ([1.0, 0.0, 0.0])\n: f --> c ([0.0, 0.0, 2.0])'
self.assertEqual(clf_str, res)
