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
|
#! /usr/bin/env python
import openturns as ot
ot.TESTPREAMBLE()
ot.PlatformInfo.SetNumericalPrecision(4)
coll1 = ot.DistributionCollection(0)
coll1.add(ot.Normal(1.0, 2.5))
coll1.add(ot.Gamma(1.5, 3.0))
pointLow = ot.Point(0)
pointLow.add(coll1[0].computeQuantile(0.25)[0])
pointLow.add(coll1[1].computeQuantile(0.25)[0])
pointHigh = ot.Point(0)
pointHigh.add(coll1[0].computeQuantile(0.75)[0])
pointHigh.add(coll1[1].computeQuantile(0.75)[0])
coll2 = ot.DistributionCollection(0)
coll2.add(ot.Gamma(2.5, 2.0))
coll2.add(ot.Normal(3.0, 1.5))
# First, check the old constructor
evaluation = ot.MarginalTransformationEvaluation(coll1)
transformation = ot.MarginalTransformationHessian(evaluation)
print("transformation=", repr(transformation))
print(
"transformation.hessian(",
repr(pointLow),
")=",
repr(transformation.hessian(pointLow)),
)
print(
"finite difference hessian(",
repr(pointLow),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(pointLow)),
)
print(
"transformation.hessian(",
repr(pointHigh),
")=",
repr(transformation.hessian(pointHigh)),
)
print(
"finite difference hessian(",
repr(pointHigh),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(pointHigh)),
)
print("input dimension=", transformation.getInputDimension())
print("output dimension=", transformation.getOutputDimension())
# Second, check the constructor for old inverse transformation
evaluation = ot.MarginalTransformationEvaluation(
coll1, ot.MarginalTransformationEvaluation.TO
)
transformation = ot.MarginalTransformationHessian(evaluation)
print("transformation=", repr(transformation))
uLow = ot.Point(coll1.getSize(), 0.25)
uHigh = ot.Point(coll1.getSize(), 0.75)
print("transformation.hessian(", repr(uLow), ")=", repr(transformation.hessian(uLow)))
print(
"finite difference hessian(",
repr(uLow),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(uLow)),
)
print("transformation.hessian(", repr(uHigh), ")=", repr(transformation.hessian(uHigh)))
print(
"finite difference hessian(",
repr(uHigh),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(uHigh)),
)
print("input dimension=", transformation.getInputDimension())
print("output dimension=", transformation.getOutputDimension())
# Third, check the constructor for the new transformation
evaluation = ot.MarginalTransformationEvaluation(coll1, coll2)
transformation = ot.MarginalTransformationHessian(evaluation)
print("transformation=", repr(transformation))
print(
"transformation.hessian(",
repr(pointLow),
")=",
repr(transformation.hessian(pointLow)),
)
print(
"finite difference hessian(",
repr(pointLow),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(pointLow)),
)
print(
"transformation.hessian(",
repr(pointHigh),
")=",
repr(transformation.hessian(pointHigh)),
)
print(
"finite difference hessian(",
repr(pointHigh),
")=",
repr(ot.CenteredFiniteDifferenceHessian(1.0e-4, evaluation).hessian(pointHigh)),
)
print("input dimension=", transformation.getInputDimension())
print("output dimension=", transformation.getOutputDimension())
|
