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
|
#! /usr/bin/env python
import openturns as ot
ot.PlatformInfo.SetNumericalPrecision(3)
distribution = ot.Beta(2.3, 2.2, -1.0, 1.0)
print(f"distribution={distribution}")
sample = distribution.getSample(1000)
factory = ot.MaximumLikelihoodFactory(ot.Beta())
infDist = factory.build(sample)
print(f"estimated distribution={infDist}")
# set (a,b) out of (r, t, a, b)
factory.setKnownParameter(
[2, 3],
[-1.0, 1.0],
)
infDist = factory.build(sample)
print(f"estimated distribution with bounds={infDist}")
factory = ot.MaximumLikelihoodFactory(ot.Exponential())
factory.setKnownParameter([1], [0.1])
print(factory.build())
print(factory.build([3, 0]))
# test BuildEstimator
distribution = ot.Gumbel(0.5, -0.5)
sample = distribution.getSample(1000)
for isRegular in [True, False]:
fResult = ot.MaximumLikelihoodFactory.BuildEstimator(
ot.GumbelFactory(), sample, isRegular
)
paramDist = fResult.getParameterDistribution()
print(f"estimator(isRegular={isRegular}): {str(paramDist)[:100]}")
|
