1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
|
# Test methods with long descriptive names can omit docstrings
# pylint: disable=missing-docstring
import unittest
import pickle
import numpy as np
import Orange
from Orange.classification import SimpleTreeLearner as SimpleTreeCls
from Orange.regression import SimpleTreeLearner as SimpleTreeReg
from Orange.data import ContinuousVariable, Domain, DiscreteVariable, Table
from Orange.tests import test_filename
class TestSimpleTreeLearner(unittest.TestCase):
def setUp(self):
self.N = 50
self.Mi = 3
self.Mf = 3
self.cls_vals = 3
np.random.seed(42)
Xi = np.random.randint(0, 2, (self.N, self.Mi)).astype(np.float64)
Xf = np.random.normal(0, 1, (self.N, self.Mf)).astype(np.float64)
X = np.hstack((Xi, Xf))
y_cls = np.random.randint(0, self.cls_vals, self.N).astype(np.float64)
y_reg = np.random.normal(0, 1, self.N).astype(np.float64)
X[np.random.random(X.shape) < 0.1] = np.nan
y_cls[np.random.random(self.N) < 0.1] = np.nan
y_reg[np.random.random(self.N) < 0.1] = np.nan
di = [Orange.data.domain.DiscreteVariable(
'd{}'.format(i), ["0", "1"]) for i in range(self.Mi)]
df = [Orange.data.domain.ContinuousVariable(
'c{}'.format(i)) for i in range(self.Mf)]
dcls = Orange.data.domain.DiscreteVariable('yc', ["0", "1", "2"])
dreg = Orange.data.domain.ContinuousVariable('yr')
domain_cls = Orange.data.domain.Domain(di + df, dcls)
domain_reg = Orange.data.domain.Domain(di + df, dreg)
self.data_cls = Orange.data.Table.from_numpy(domain_cls, X, y_cls)
self.data_reg = Orange.data.Table.from_numpy(domain_reg, X, y_reg)
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_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_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_regression(self):
lrn = SimpleTreeReg()
clf = lrn(self.data_reg)
p = clf(self.data_reg)
self.assertEqual(p.shape, (self.N,))
def test_SimpleTree_regression_pickle(self):
pass
def test_SimpleTree_regression_tree(self):
lrn = SimpleTreeReg(min_instances=5)
clf = lrn(self.data_reg)
self.assertEqual(
clf.dumps_tree(clf.node),
'{ 0 2 { 1 4 0.13895 { 1 4 -0.32607 { 2 4.60993 1.71141 } '
'{ 2 4.96454 3.56122 } } { 2 7.09220 -4.32343 } } { 1 4 -0.35941 '
'{ 0 0 { 1 5 -0.20027 { 2 3.54255 0.95095 } { 2 5.50000 -5.56049 } '
'} { 2 7.62411 2.03615 } } { 1 5 0.40797 { 1 3 0.83459 '
'{ 2 3.71094 0.27028 } { 2 5.18490 3.70920 } } { 2 5.77083 5.93398 '
'} } } }')
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), 1)
def test_SimpleTree_to_string_classification(self):
domain = Domain([DiscreteVariable(name='d1', values='ef'),
ContinuousVariable(name='c1')],
DiscreteVariable(name='cls', values='abc'))
data = Table.from_list(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 = '\n' \
'd1 ([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)
def test_SimpleTree_to_string_regression(self):
domain = Domain([DiscreteVariable(name='d1', values='ef'),
ContinuousVariable(name='c1')],
ContinuousVariable(name='cls'))
data = Table.from_list(domain, [['e', 1, 10],
['e', 1, 20],
['e', 2, 20],
['f', 2, 30],
["e", 3, 10],
['f', 3, 30]])
lrn = SimpleTreeReg(min_instances=1)
reg = lrn(data)
reg_str = reg.to_string()
res = '\n' \
'd1 (20: 6.0)\n' \
': e\n' \
' c1 (15: 4.0)\n' \
' : <=2.5\n' \
' c1 (16.6667: 3.0)\n' \
' : <=1.5 --> (15: 2.0)\n' \
' : >1.5 --> (20: 1.0)\n' \
' : >2.5 --> (10: 1.0)\n' \
': f --> (30: 2.0)'
self.assertEqual(reg_str, res)
def test_SimpleTree_to_string_cls_decimals(self):
data = Table(test_filename("datasets/lenses.tab"))
lrn = SimpleTreeReg(min_instances=1)
cls = lrn(data)
cls_str = cls.to_string()
res = ' astigmatic ([4.0, 3.0, 5.0])'
self.assertEqual(cls_str.split("\n")[3], res)
def test_SimpleTree_to_string_reg_decimals(self):
data = Table("housing")
lrn = SimpleTreeReg(min_instances=1)
reg = lrn(data)
reg_str = reg.to_string()
res = ' LSTAT (19.9: 430.0)'
self.assertEqual(reg_str.split("\n")[3], res)
if __name__ == '__main__':
unittest.main()
|
