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
|
import unittest
import Tasmanian
DREAM = Tasmanian.DREAM
import os
import numpy as np
import testCommon
ttc = testCommon.TestTasCommon()
class TestTasClass(unittest.TestCase):
'''
Tests for the Tasmanian DREAM python bindings module
'''
def __init__(self):
unittest.TestCase.__init__(self, "testNothing")
def testNothing(self):
pass
def checkGaussian3D(self):
iNumDimensions = 3
iNumSamples = 1000
iNumChains = 20
iNumIterations = int(iNumSamples / iNumChains) + 2
iNumBurnup = iNumIterations
state = DREAM.State(iNumChains, iNumDimensions)
state.setState(DREAM.genGaussianSamples([2.0, 2.0, 2.0], [3.0, 3.0, 3.0], iNumChains, DREAM.RandomGenerator("minstd_rand", 42)))
DREAM.Sample(iNumBurnup, iNumSamples,
lambda x : np.exp( -0.5 * np.sum( (x - 2.0)**2, 1 ) / 9.0 ), # Gaussian mean 2.0 variance 9.0
DREAM.Domain("unbounded"),
state,
DREAM.IndependentUpdate("gaussian", 3.0),
DREAM.DifferentialUpdate(0),
DREAM.RandomGenerator("minstd_rand", 42))
aMean, aVar = state.getHistoryMeanVariance()
self.assertTrue(np.max(np.abs(aMean - 2.0)) < 0.2, "Failed the mean test for Gaussian 3D")
self.assertTrue(np.max(np.abs(aVar - 9.0)) < 1.0, "Failed the variance test for Gaussian 3D")
state = DREAM.State(iNumChains, iNumDimensions)
state.setState(DREAM.genGaussianSamples([1.5, 2.0, 2.5], [0.5, 1.0, 2.0], iNumChains, DREAM.RandomGenerator("minstd_rand", 42)))
likely = DREAM.LikelihoodGaussAnisotropic(np.array([0.25, 1.0, 4.0]), np.array([1.5, 2.0, 2.5]), 1)
DREAM.Sample(iNumBurnup, iNumSamples,
lambda x : likely.getLikelihood(DREAM.typeLogform, x), # Gaussian mean 2.0 variance 9.0
DREAM.Domain("unbounded"),
state,
DREAM.IndependentUpdate("gaussian", 3.0),
DREAM.DifferentialUpdate(97),
DREAM.RandomGenerator("minstd_rand", 72),
DREAM.typeLogform)
aMean, aVar = state.getHistoryMeanVariance()
self.assertTrue(np.max(np.abs(aMean - np.array([1.5, 2.0, 2.5]))) < 0.4, "Failed the mean test for anisotropic Gaussian 3D")
self.assertTrue(np.max(np.abs(aVar - np.array([0.25, 1.0, 4.0]))) < 1.0, "Failed the variance test for anisotropic Gaussian 3D")
def checkGrid(self):
iNumDimensions = 2
iNumSamples = 1000
iNumChains = 20
iNumIterations = int(iNumSamples / iNumChains) + 2
iNumBurnup = iNumIterations
grid = Tasmanian.SparseGrid()
grid.makeSequenceGrid(2, 1, 24, "level", "rleja")
Tasmanian.loadNeededPoints(lambda x, tid : np.ones((1,)) * np.exp( -0.5 * np.sum((x - 0.3)**2) / 0.01 ), grid, 1)
state = DREAM.State(iNumChains, grid)
state.setState(DREAM.genGaussianSamples([0.3, 0.3], [0.1, 0.1], iNumChains, DREAM.RandomGenerator("minstd_rand", 55)))
DREAM.Sample(iNumBurnup, iNumSamples,
DREAM.PosteriorEmbeddedLikelihood(grid, "uniform"),
DREAM.Domain("hypercube", np.array((0.0, 0.0)), np.array((1.0, 1.0))),
state,
DREAM.IndependentUpdate("uniform", 0.1),
DREAM.DifferentialUpdate(98),
DREAM.RandomGenerator("minstd_rand", 33))
aMean, aVar = state.getHistoryMeanVariance()
self.assertTrue(np.max(np.abs(aMean - np.array([0.3, 0.3]))) < 0.05, "Failed the mean test for grid likelihood")
self.assertTrue(np.max(np.abs(aVar - np.array([0.01, 0.01]))) < 0.005, "Failed the variance test for grid likelihood")
def checkModel(self):
iNumDimensions = 2
iNumSamples = 1000
iNumChains = 20
iNumIterations = int(iNumSamples / iNumChains) + 2
iNumBurnup = iNumIterations
likely = DREAM.LikelihoodGaussIsotropic(0.01, np.array([1.0, 2.0]))
state = DREAM.State(iNumChains, iNumDimensions)
state.setState(DREAM.genUniformSamples([-1.0, -1.0], [1.0, 1.0], iNumChains, DREAM.RandomGenerator("minstd_rand", 67)))
DREAM.Sample(iNumBurnup, iNumSamples,
DREAM.Posterior(lambda x : np.matmul(np.cos(x), np.array([[1.0, 0.0], [0.0, 2.0]])),
likely, "uniform", DREAM.typeLogform),
DREAM.Domain("hypercube", np.array((0.0, 0.0)), np.array((1.0, 1.0))),
state,
DREAM.IndependentUpdate("uniform", 0.1),
DREAM.DifferentialUpdate(98),
DREAM.RandomGenerator("minstd_rand", 88),
DREAM.typeLogform)
aMode = state.getApproximateMode()
self.assertTrue(np.max(np.abs(aMode)) < 0.01, "Failed the mode test for custom model and Tasmanian likelihood")
def performDreamTests(self):
self.checkGaussian3D() # analytic model/likelihood
self.checkGrid()
self.checkModel() # custom model
if __name__ == "__main__":
tester = TestTasClass()
tester.performDreamTests()
|
