|
1
|
|
|
import unittest |
|
2
|
|
|
from random import random |
|
3
|
|
|
|
|
4
|
|
|
from numpy.random.mtrand import normal |
|
5
|
|
|
|
|
6
|
|
|
from pysie.dsl.two_groups import MeanDiffTesting, ProportionDiffTesting |
|
7
|
|
|
from pysie.stats.distributions import MeanDiffSamplingDistribution, DistributionFamily, \ |
|
8
|
|
|
ProportionDiffSamplingDistribution |
|
9
|
|
|
from pysie.stats.samples import Sample, SampleDistribution |
|
10
|
|
|
|
|
11
|
|
|
|
|
12
|
|
|
class MeanDiffTestingUnitTest(unittest.TestCase): |
|
13
|
|
|
|
|
14
|
|
View Code Duplication |
def test_normal(self): |
|
|
|
|
|
|
15
|
|
|
grp1_mu = 0.0 |
|
16
|
|
|
grp1_sigma = 1.0 |
|
17
|
|
|
grp1_sample_size = 31 |
|
18
|
|
|
grp1_sample = Sample() |
|
19
|
|
|
|
|
20
|
|
|
grp2_mu = 0.09 |
|
21
|
|
|
grp2_sigma = 2.0 |
|
22
|
|
|
grp2_sample_size = 36 |
|
23
|
|
|
grp2_sample = Sample() |
|
24
|
|
|
|
|
25
|
|
|
for i in range(grp1_sample_size): |
|
26
|
|
|
grp1_sample.add_numeric(normal(grp1_mu, grp1_sigma)) |
|
27
|
|
|
|
|
28
|
|
|
for i in range(grp2_sample_size): |
|
29
|
|
|
grp2_sample.add_numeric(normal(grp2_mu, grp2_sigma)) |
|
30
|
|
|
|
|
31
|
|
|
sampling_distribution = MeanDiffSamplingDistribution(grp1_sample_distribution=SampleDistribution(grp1_sample), |
|
32
|
|
|
grp2_sample_distribution=SampleDistribution(grp2_sample)) |
|
33
|
|
|
self.assertEqual(sampling_distribution.distribution_family, DistributionFamily.normal) |
|
34
|
|
|
testing = MeanDiffTesting(sampling_distribution=sampling_distribution) |
|
35
|
|
|
print('one tail p-value: ' + str(testing.p_value_one_tail)) |
|
36
|
|
|
print('two tail p-value: ' + str(testing.p_value_two_tail)) |
|
37
|
|
|
reject_one_tail, reject_two_tail = testing.will_reject(0.01) |
|
38
|
|
|
print('will reject mean_1 == mean_2 (one-tail) ? ' + str(reject_one_tail)) |
|
39
|
|
|
print('will reject mean_1 == mean_2 (two-tail) ? ' + str(reject_two_tail)) |
|
40
|
|
|
self.assertFalse(reject_one_tail) |
|
41
|
|
|
self.assertFalse(reject_two_tail) |
|
42
|
|
|
|
|
43
|
|
View Code Duplication |
def test_student(self): |
|
|
|
|
|
|
44
|
|
|
grp1_mu = 0.0 |
|
45
|
|
|
grp1_sigma = 1.0 |
|
46
|
|
|
grp1_sample_size = 29 |
|
47
|
|
|
grp1_sample = Sample() |
|
48
|
|
|
|
|
49
|
|
|
grp2_mu = 0.09 |
|
50
|
|
|
grp2_sigma = 2.0 |
|
51
|
|
|
grp2_sample_size = 28 |
|
52
|
|
|
grp2_sample = Sample() |
|
53
|
|
|
|
|
54
|
|
|
for i in range(grp1_sample_size): |
|
55
|
|
|
grp1_sample.add_numeric(normal(grp1_mu, grp1_sigma)) |
|
56
|
|
|
|
|
57
|
|
|
for i in range(grp2_sample_size): |
|
58
|
|
|
grp2_sample.add_numeric(normal(grp2_mu, grp2_sigma)) |
|
59
|
|
|
|
|
60
|
|
|
sampling_distribution = MeanDiffSamplingDistribution(grp1_sample_distribution=SampleDistribution(grp1_sample), |
|
61
|
|
|
grp2_sample_distribution=SampleDistribution(grp2_sample)) |
|
62
|
|
|
self.assertEqual(sampling_distribution.distribution_family, DistributionFamily.student_t) |
|
63
|
|
|
testing = MeanDiffTesting(sampling_distribution=sampling_distribution) |
|
64
|
|
|
print('one tail p-value: ' + str(testing.p_value_one_tail)) |
|
65
|
|
|
print('two tail p-value: ' + str(testing.p_value_two_tail)) |
|
66
|
|
|
reject_one_tail, reject_two_tail = testing.will_reject(0.01) |
|
67
|
|
|
print('will reject mean_1 == mean_2 (one-tail) ? ' + str(reject_one_tail)) |
|
68
|
|
|
print('will reject mean_1 == mean_2 (two-tail) ? ' + str(reject_two_tail)) |
|
69
|
|
|
self.assertFalse(reject_one_tail) |
|
70
|
|
|
self.assertFalse(reject_two_tail) |
|
71
|
|
|
|
|
72
|
|
|
|
|
73
|
|
|
class ProportionDiffTestingUnitTest(unittest.TestCase): |
|
74
|
|
|
|
|
75
|
|
View Code Duplication |
def test_normal(self): |
|
|
|
|
|
|
76
|
|
|
grp1_sample = Sample() |
|
77
|
|
|
grp2_sample = Sample() |
|
78
|
|
|
|
|
79
|
|
|
for i in range(100): |
|
80
|
|
|
if random() <= 0.6: |
|
81
|
|
|
grp1_sample.add_category("OK") |
|
82
|
|
|
else: |
|
83
|
|
|
grp1_sample.add_category("CANCEL") |
|
84
|
|
|
|
|
85
|
|
|
for i in range(100): |
|
86
|
|
|
if random() <= 0.61: |
|
87
|
|
|
grp2_sample.add_category("OK") |
|
88
|
|
|
else: |
|
89
|
|
|
grp2_sample.add_category("CANCEL") |
|
90
|
|
|
|
|
91
|
|
|
sampling_distribution = ProportionDiffSamplingDistribution(grp1_sample_distribution=SampleDistribution( |
|
92
|
|
|
grp1_sample, categorical_value="OK"), |
|
93
|
|
|
grp2_sample_distribution=SampleDistribution( |
|
94
|
|
|
grp2_sample, categorical_value="OK")) |
|
95
|
|
|
self.assertEqual(sampling_distribution.distribution_family, DistributionFamily.normal) |
|
96
|
|
|
|
|
97
|
|
|
testing = ProportionDiffTesting(sampling_distribution=sampling_distribution) |
|
98
|
|
|
print('one tail p-value: ' + str(testing.p_value_one_tail)) |
|
99
|
|
|
print('two tail p-value: ' + str(testing.p_value_two_tail)) |
|
100
|
|
|
reject_one_tail, reject_two_tail = testing.will_reject(0.01) |
|
101
|
|
|
print('will reject p_1 == p_2 (one-tail) ? ' + str(reject_one_tail)) |
|
102
|
|
|
print('will reject p_1 == p_2 (two-tail) ? ' + str(reject_two_tail)) |
|
103
|
|
|
self.assertFalse(reject_one_tail) |
|
104
|
|
|
self.assertFalse(reject_two_tail) |
|
105
|
|
|
|
|
106
|
|
View Code Duplication |
def test_student(self): |
|
|
|
|
|
|
107
|
|
|
grp1_sample = Sample() |
|
108
|
|
|
grp2_sample = Sample() |
|
109
|
|
|
|
|
110
|
|
|
for i in range(20): |
|
111
|
|
|
if random() <= 0.6: |
|
112
|
|
|
grp1_sample.add_category("OK") |
|
113
|
|
|
else: |
|
114
|
|
|
grp1_sample.add_category("CANCEL") |
|
115
|
|
|
|
|
116
|
|
|
for i in range(20): |
|
117
|
|
|
if random() <= 0.61: |
|
118
|
|
|
grp2_sample.add_category("OK") |
|
119
|
|
|
else: |
|
120
|
|
|
grp2_sample.add_category("CANCEL") |
|
121
|
|
|
|
|
122
|
|
|
sampling_distribution = ProportionDiffSamplingDistribution(grp1_sample_distribution=SampleDistribution( |
|
123
|
|
|
grp1_sample, categorical_value="OK"), |
|
124
|
|
|
grp2_sample_distribution=SampleDistribution( |
|
125
|
|
|
grp2_sample, categorical_value="OK")) |
|
126
|
|
|
self.assertEqual(sampling_distribution.distribution_family, DistributionFamily.simulation) |
|
127
|
|
|
|
|
128
|
|
|
testing = ProportionDiffTesting(sampling_distribution=sampling_distribution) |
|
129
|
|
|
print('one tail p-value: ' + str(testing.p_value_one_tail)) |
|
130
|
|
|
print('two tail p-value: ' + str(testing.p_value_two_tail)) |
|
131
|
|
|
reject_one_tail, reject_two_tail = testing.will_reject(0.01) |
|
132
|
|
|
print('will reject p_1 == p_2 (one-tail) ? ' + str(reject_one_tail)) |
|
133
|
|
|
print('will reject p_1 == p_2 (two-tail) ? ' + str(reject_two_tail)) |
|
134
|
|
|
self.assertFalse(reject_one_tail) |
|
135
|
|
|
self.assertFalse(reject_two_tail) |
|
136
|
|
|
|
|
137
|
|
|
if __name__ == '__main__': |
|
138
|
|
|
unittest.main() |
chen0040 /
pysie
Loading history...