def test_fit_2hp(self):
# carry out fit
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
T = bl.tm.CombinedTransitionModel(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 2),
target='mean'),
bl.tm.RegimeSwitch('log10pMin', [-3, -1]))
S.setTM(T)
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5], [
5.80970506e-03, 1.12927905e-01, 4.44501254e-02, 1.00250119e-02,
1.72751309e-05
],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[0.96492471, 2.09944204, 2.82451616, 3.72702495, 5.0219119],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-10.7601875492,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]), [[0., 0.2], [0.48943645, 0.51056355]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
# test joint hyper-parameter distribution
x, y, p = S.getJointHyperParameterDistribution(['log10pMin', 'sigma'])
np.testing.assert_allclose(
np.array([x, y]), [[-3., -1.], [0., 0.2]],
rtol=1e-05,
err_msg='Erroneous parameter values in joint hyper-parameter '
'distribution.')
np.testing.assert_allclose(
p, [[0.00701834, 0.0075608], [0.48241812, 0.50300274]],
rtol=1e-05,
err_msg='Erroneous probability values in joint hyper-parameter '
'distribution.')
def test_fit_1cp_1bp_2hp(self):
# carry out fit
S = bl.ChangepointStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
T = bl.tm.SerialTransitionModel(
bl.tm.Static(), bl.tm.ChangePoint('ChangePoint', [0, 1]),
bl.tm.CombinedTransitionModel(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 2),
target='mean'),
bl.tm.RegimeSwitch('log10pMin', [-3, -1])),
bl.tm.BreakPoint('BreakPoint', 'all'), bl.tm.Static())
S.setTM(T)
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.01243717, 0.03016095, 0.016939, 0.00024909, 0.00024909],
rtol=1e-04,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[0.96802204, 1.95705078, 3.47078681, 4.22225665, 4.22225665],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-15.072007461556161,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]), [[0., 0.2], [0.4963324, 0.5036676]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
# test duration distribution
d, p = S.getDurationDistribution(['ChangePoint', 'BreakPoint'])
np.testing.assert_allclose(
np.array([d, p]),
[[1., 2., 3.], [0.01039273, 0.49395867, 0.49564861]],
rtol=1e-05,
err_msg='Erroneous values in duration distribution.')
def test_step_add2TM_2hp_prior_hyperpriors_TMprior(self):
# carry out fit
S = bl.OnlineStudy(storeHistory=True)
S.setOM(bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma', bl.oint(0, 2, 20), prior=lambda m, s: 1./s))
T1 = bl.tm.CombinedTransitionModel(bl.tm.GaussianRandomWalk('s1', [0.25, 0.5],
target='mean',
prior=stats.Exponential('e', 0.5)),
bl.tm.GaussianRandomWalk('s2', bl.cint(0, 0.2, 2),
target='sigma',
prior=np.array([0.2, 0.8]))
)
T2 = bl.tm.Independent()
S.addTransitionModel('T1', T1)
S.addTransitionModel('T2', T2)
S.setTransitionModelPrior([0.9, 0.1])
data = np.array([1, 2, 3, 4, 5])
for d in data:
S.step(d)
# test transition model distributions
np.testing.assert_allclose(S.getCurrentTransitionModelDistribution(local=False)[1],
[0.49402616, 0.50597384],
rtol=1e-05, err_msg='Erroneous transition model probabilities.')
np.testing.assert_allclose(S.getCurrentTransitionModelDistribution(local=True)[1],
[0.81739495, 0.18260505],
rtol=1e-05, err_msg='Erroneous local transition model probabilities.')
# test hyper-parameter distributions
np.testing.assert_allclose(S.getCurrentHyperParameterDistribution('s2')[1],
[0.19047162, 0.80952838],
rtol=1e-05, err_msg='Erroneous hyper-parameter distribution.')
# test parameter distributions
np.testing.assert_allclose(S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.05825921, 0.20129444, 0.07273516, 0.02125759, 0.0039255],
rtol=1e-05, err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(S.getParameterMeanValues('mean'),
[1.0771838, 1.71494272, 2.45992376, 3.34160617, 4.39337253],
rtol=1e-05, err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence, -9.46900822686, decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_prior_sympy(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian(
'mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=[stats.Uniform('u', 0, 6),
stats.Exponential('e', 2.)]))
S.setTM(bl.tm.GaussianRandomWalk('sigma', 0.1, target='mean'))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.00909976, 0.0089861, 0.00887967, 0.00881235, 0.00880499],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.9942575, 2.99646768, 3., 3.00353232, 3.0057425],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-12.4324853153,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_2hp(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
T = bl.tm.CombinedTransitionModel(
bl.tm.GaussianRandomWalk('sigma', 0.1, target='mean'),
bl.tm.RegimeSwitch('log10pMin', -3))
S.setTM(T)
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.02976422, 0.15404218, 0.10859567, 0.02553673, 0.00054109],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[1.08288559, 2.24388932, 2.38033179, 2.98934128, 4.64547841],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-14.3305753098,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_0hp(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
S.setTM(bl.tm.Static())
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.00707902, 0.00707902, 0.00707902, 0.00707902, 0.00707902],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(S.getParameterMeanValues('mean'),
[3., 3., 3., 3., 3.],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-16.1946904707,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_1hp(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
S.setTM(bl.tm.GaussianRandomWalk('sigma', 0.1, target='mean'))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.00722368, 0.00712209, 0.00702789, 0.00696926, 0.00696322],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.99313985, 2.99573566, 3., 3.00426434, 3.00686015],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-16.1865343702,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_prior_array(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=np.ones((20, 20))))
S.setTM(bl.tm.GaussianRandomWalk('sigma', 0.1, target='mean'))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.04317995, 0.04296549, 0.04275526, 0.04262151, 0.04262491],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.66415455, 2.66519273, 2.66664847, 2.66788051, 2.66828383],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-10.9827282104,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_prior_function(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1. / s))
S.setTM(bl.tm.GaussianRandomWalk('sigma', 0.1, target='mean'))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.01591204, 0.01579036, 0.01567361, 0.01559665, 0.01558591],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.99576496, 2.99741879, 3., 3.00258121, 3.00423504],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-11.9842221343,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_optimize(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma', bl.oint(0, 2, 20)))
T = bl.tm.CombinedTransitionModel(bl.tm.GaussianRandomWalk('sigma', 1.07, target='mean'),
bl.tm.RegimeSwitch('log10pMin', -3.90))
S.setTM(T)
S.optimize()
# test parameter distributions
np.testing.assert_allclose(S.getParameterDistributions('mean', density=False)[1][:, 5],
[4.52572851e-04, 1.67790320e-03, 2.94525791e-07, 1.49841548e-08, 1.10238422e-09],
rtol=1e-05, err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(S.getParameterMeanValues('mean'),
[0.95899404, 1.93816557, 2.99999968, 4.06183394, 5.04100612],
rtol=1e-05, err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence, -8.01030368139, decimal=5,
err_msg='Erroneous log-evidence value.')
# test optimized hyper-parameter values
np.testing.assert_almost_equal(S.getHyperParameterValue('sigma'), 1.06576569677, decimal=5,
err_msg='Erroneous log-evidence value.')
np.testing.assert_almost_equal(S.getHyperParameterValue('log10pMin'), -4.04001476542, decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_0hp(self):
# carry out fit (this test is designed to fall back on the fit method of the Study class)
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
S.setTM(bl.tm.Static())
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.00707902, 0.00707902, 0.00707902, 0.00707902, 0.00707902],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(S.getParameterMeanValues('mean'),
[3., 3., 3., 3., 3.],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-16.1946904707,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_step_set1TM_0hp(self):
# carry out fit
S = bl.OnlineStudy(storeHistory=True)
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
S.setTM(bl.tm.Static())
data = np.array([1, 2, 3, 4, 5])
for d in data:
S.step(d)
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean')[1][:, 5],
[0.0053811, 0.38690331, 0.16329865, 0.04887604, 0.01334921],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[0.96310103, 1.5065597, 2.00218465, 2.500366, 3.],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-16.1946904707,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_1hp(self):
# carry out fit
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
S.setTM(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 2),
target='mean'))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.01042107, 0.00766233, 0.00618352, 0.00554651, 0.00548637],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.88534505, 2.93135361, 3., 3.06864639, 3.11465495],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-16.0629517262,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
print(np.array([x, p]))
np.testing.assert_allclose(
np.array([x, p]), [[0., 0.2], [0.43828499, 0.56171501]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
def test_fit_hyperprior_function(self):
# carry out fit
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
S.setTM(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0.1, 0.3, 2),
target='mean',
prior=lambda s: 1. / s))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.04071021, 0.00783661, 0.00527211, 0.00484169, 0.00480379],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.68460027, 2.81872578, 3., 3.18127422, 3.31539973],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-15.9898700147,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]), [[0.1, 0.3], [0.61609973, 0.38390027]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
def test_fit_hyperprior_array(self):
# carry out fit
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
S.setTM(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 2),
target='mean',
prior=np.array([0.2, 0.8])))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.01205759, 0.00794796, 0.00574501, 0.00479608, 0.00470649],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.82920111, 2.89773902, 3., 3.10226098, 3.17079889],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-15.9915077133,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]), [[0., 0.2], [0.16322581, 0.83677419]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
def test_fit_hyperprior_sympy(self):
# carry out fit
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
S.setTM(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 2),
target='mean',
prior=stats.Exponential('e', 1.)))
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.01012545, 0.00761074, 0.00626273, 0.00568207, 0.00562725],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[2.89548676, 2.93742566, 3., 3.06257434, 3.10451324],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-17.0866290887,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]), [[0., 0.2], [0.487971, 0.512029]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
def test_save_load(self):
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean', bl.cint(0, 6, 20), 'sigma',
bl.oint(0, 2, 20)))
S.setTM(bl.tm.Static())
S.fit()
bl.save('study.bl', S)
S = bl.load('study.bl')
def test_dynamichyperparameter(self):
S = bl.OnlineStudy(storeHistory=True)
S.setOM(bl.om.Poisson('rate', bl.oint(0, 6, 50)))
S.add(
'gradual',
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 5),
target='rate'))
S.add('static', bl.tm.Static())
for d in np.arange(5):
S.step(d)
p = S.eval('exp(0.99*log(sigma@2))+1 > 1.1')
np.testing.assert_almost_equal(
p,
0.61228433813735061,
decimal=5,
err_msg='Erroneous parsing result for inequality.')
S = bl.OnlineStudy(storeHistory=False)
S.setOM(bl.om.Poisson('rate', bl.oint(0, 6, 50)))
S.add(
'gradual',
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 5),
target='rate'))
S.add('static', bl.tm.Static())
for d in np.arange(3):
S.step(d)
p = S.eval('exp(0.99*log(sigma))+1 > 1.1')
np.testing.assert_almost_equal(
p,
0.61228433813735061,
decimal=5,
err_msg='Erroneous parsing result for inequality.')
def test_scipy_2p(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
L = bl.om.SciPy(scipy.stats.norm, 'loc', bl.cint(0, 7, 200), 'scale', bl.oint(0, 1, 200))
S.setOM(L)
S.setTM(bl.tm.Static())
S.fit()
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence, -13.663836264357225, decimal=5,
err_msg='Erroneous log-evidence value.')
def test_optimize(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
T = bl.tm.CombinedTransitionModel(
bl.tm.GaussianRandomWalk('sigma', 1.07, target='mean'),
bl.tm.RegimeSwitch('log10pMin', -3.90))
S.setTM(T)
S.optimize()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5], [
4.525547e-04, 1.677968e-03, 2.946498e-07, 1.499508e-08,
1.102637e-09
],
rtol=1e-05,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[0.95899404, 1.93816557, 2.99999968, 4.06183394, 5.04100612],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-8.010466752050611,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test optimized hyper-parameter values
np.testing.assert_almost_equal(S.getHyperParameterValue('sigma'),
1.065854087589326,
decimal=5,
err_msg='Erroneous log-evidence value.')
np.testing.assert_almost_equal(S.getHyperParameterValue('log10pMin'),
-4.039735868499399,
decimal=5,
err_msg='Erroneous log-evidence value.')
def test_statichyperparameter(self):
S = bl.HyperStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(bl.om.Poisson('rate', bl.oint(0, 6, 50)))
S.setTM(
bl.tm.GaussianRandomWalk('sigma',
bl.cint(0, 0.2, 5),
target='rate'))
S.fit()
p = S.eval('exp(0.99*log(sigma))+1 > 1.1')
np.testing.assert_almost_equal(
p,
0.60696006616644793,
decimal=5,
err_msg='Erroneous parsing result for inequality.')
def test_sympy_2p(self):
# carry out fit
S = bl.Study()
S.loadData(np.array([1, 2, 3, 4, 5]))
mu = Symbol('mu')
std = Symbol('std', positive=True)
normal = sympy.stats.Normal('norm', mu, std)
L = bl.om.SymPy(normal, 'mu', bl.cint(0, 7, 200), 'std', bl.oint(0, 1, 200), prior=lambda x, y: 1.)
S.setOM(L)
S.setTM(bl.tm.Static())
S.fit()
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence, -13.663836264357226, decimal=5,
err_msg='Erroneous log-evidence value.')
def test_fit_hyperpriors(self):
# carry out fit
S = bl.ChangepointStudy()
S.loadData(np.array([1, 2, 3, 4, 5]))
S.setOM(
bl.om.Gaussian('mean',
bl.cint(0, 6, 20),
'sigma',
bl.oint(0, 2, 20),
prior=lambda m, s: 1 / s**3))
T = bl.tm.SerialTransitionModel(
bl.tm.Static(),
bl.tm.ChangePoint('ChangePoint', [0, 1],
prior=np.array([0.3, 0.7])),
bl.tm.CombinedTransitionModel(
bl.tm.GaussianRandomWalk('sigma',
bl.oint(0, 0.2, 2),
target='mean',
prior=lambda s: 1. / s),
bl.tm.RegimeSwitch('log10pMin', [-3, -1])),
bl.tm.BreakPoint('BreakPoint',
'all',
prior=stats.Normal('Normal', 3., 1.)),
bl.tm.Static())
S.setTM(T)
S.fit()
# test parameter distributions
np.testing.assert_allclose(
S.getParameterDistributions('mean', density=False)[1][:, 5],
[0.03372851, 0.05087598, 0.02024129, 0.00020918, 0.00020918],
rtol=1e-04,
err_msg='Erroneous posterior distribution values.')
# test parameter mean values
np.testing.assert_allclose(
S.getParameterMeanValues('mean'),
[0.9894398, 1.92805399, 3.33966456, 4.28759449, 4.28759449],
rtol=1e-05,
err_msg='Erroneous posterior mean values.')
# test model evidence value
np.testing.assert_almost_equal(S.logEvidence,
-15.709534690217343,
decimal=5,
err_msg='Erroneous log-evidence value.')
# test hyper-parameter distribution
x, p = S.getHyperParameterDistribution('sigma')
np.testing.assert_allclose(
np.array([x, p]),
[[0.06666667, 0.13333333], [0.66515107, 0.33484893]],
rtol=1e-05,
err_msg='Erroneous values in hyper-parameter distribution.')
# test duration distribution
d, p = S.getDurationDistribution(['ChangePoint', 'BreakPoint'])
np.testing.assert_allclose(
np.array([d, p]),
[[1., 2., 3.], [0.00373717, 0.40402616, 0.59223667]],
rtol=1e-05,
err_msg='Erroneous values in duration distribution.')
