def test_gp_nonfinite_phenotype():
random = RandomState(94584)
N = 50
X = random.randn(N, 100)
offset = 0.5
mean = OffsetMean(N)
mean.offset = offset
mean.fix_offset()
cov = LinearCov(X)
cov.scale = 1.0
y = zeros(N)
y[0] = nan
with pytest.raises(ValueError):
GP(y, mean, cov)
y[0] = -inf
with pytest.raises(ValueError):
GP(y, mean, cov)
y[0] = +inf
with pytest.raises(ValueError):
GP(y, mean, cov)
def _get_data():
random = RandomState(0)
N = 10
X = random.randn(N, N + 1)
X -= X.mean(0)
X /= X.std(0)
X /= sqrt(X.shape[1])
offset = 1.0
mean = OffsetMean(N)
mean.offset = offset
cov_left = LinearCov(X)
cov_left.scale = 1.5
cov_right = EyeCov(N)
cov_right.scale = 1.5
cov = SumCov([cov_left, cov_right])
lik = BernoulliProdLik(LogitLink())
y = GGPSampler(lik, mean, cov).sample(random)
return dict(mean=mean,
cov=cov,
lik=lik,
y=y,
cov_left=cov_left,
cov_right=cov_right)
def _outcome_sample(random, offset, X):
n = X.shape[0]
mean = OffsetMean(n)
mean.offset = offset
cov_left = LinearCov(X)
cov_left.scale = 1.5
cov_right = EyeCov(n)
cov_right.scale = 1.5
cov = SumCov([cov_left, cov_right])
lik = DeltaProdLik()
return GGPSampler(lik, mean, cov).sample(random)
def test_gp_value_1():
random = RandomState(94584)
N = 50
X = random.randn(N, 100)
offset = 0.5
mean = OffsetMean(N)
mean.offset = offset
mean.fix_offset()
cov = LinearCov(X)
cov.scale = 1.0
y = random.randn(N)
gp = GP(y, mean, cov)
assert_allclose(gp.value(), -153.623791551399108)
def test_gp_gradient():
random = RandomState(94584)
N = 50
X = random.randn(N, 100)
offset = 0.5
mean = OffsetMean(N)
mean.offset = offset
mean.fix_offset()
cov = LinearCov(X)
cov.scale = 1.0
y = random.randn(N)
gp = GP(y, mean, cov)
assert_allclose(gp._check_grad(), 0, atol=1e-5)
def test_gp_value_2():
random = RandomState(94584)
N = 50
X1 = random.randn(N, 3)
X2 = random.randn(N, 100)
mean = LinearMean(X1)
cov = LinearCov(X2)
cov.scale = 1.0
y = random.randn(N)
gp = GP(y, mean, cov)
assert_allclose(gp.value(), -153.091074766)
mean.effsizes = [3.4, 1.11, -6.1]
assert_allclose(gp.value(), -178.273116338)
def test_sumcov():
random = np.random.RandomState(0)
X = random.randn(3, 2)
cov_left = LinearCov(X)
K = random.randn(3, 3)
K = K @ K.T
cov_right = GivenCov(K)
cov = SumCov([cov_left, cov_right])
assert_allclose(cov.value(), cov_left.value() + cov_right.value())
assert_allclose(cov._check_grad(), 0, atol=1e-5)
cov_left.scale = 0.1
assert_allclose(cov._check_grad(), 0, atol=1e-5)
with pytest.raises(ValueError):
K = random.randn(3, 3)
GivenCov(K)
def linear_eye_cov():
from numpy import sqrt
from numpy.random import RandomState
from glimix_core.cov import EyeCov, LinearCov, SumCov
random = RandomState(458)
X = random.randn(_nsamples(), _nsamples() + 1)
X -= X.mean(0)
X /= X.std(0)
X /= sqrt(X.shape[1])
cov_left = LinearCov(X)
cov_left.scale = 1.0
cov_right = EyeCov(_nsamples())
cov_right.scale = 1.0
return SumCov([cov_left, cov_right])
def test_gp_maximize():
random = RandomState(94584)
N = 50
X = random.randn(N, 100)
offset = 0.5
mean = OffsetMean(N)
mean.offset = offset
mean.fix_offset()
cov = LinearCov(X)
cov.scale = 1.0
y = random.randn(N)
gp = GP(y, mean, cov)
assert_allclose(gp.value(), -153.623791551)
gp.fit(verbose=False)
assert_allclose(gp.value(), -79.8992122415)
def test_linearcov():
X = RandomState(0).randn(3, 2)
cov = LinearCov(X)
assert_allclose(cov.value(), X @ X.T)
assert_allclose(cov._check_grad(), 0, atol=1e-5)
assert_allclose(cov.scale, 1.0)
cov.scale = 1.5
assert_allclose(cov.scale, 1.5)
assert_allclose(cov.value(), 1.5 * (X @ X.T))
assert_allclose(cov._check_grad(), 0, atol=1e-5)
def test_lmm_predict():
random = RandomState(9458)
n = 30
X = random.randn(n, n + 1)
X -= X.mean(0)
X /= X.std(0)
X /= sqrt(X.shape[1])
offset = 1.0
mean = OffsetMean(n)
mean.offset = offset
cov_left = LinearCov(X)
cov_left.scale = 1.5
cov_right = EyeCov(n)
cov_right.scale = 1.5
cov = SumCov([cov_left, cov_right])
lik = DeltaProdLik()
y = GGPSampler(lik, mean, cov).sample(random)
QS = economic_qs_linear(X)
lmm = LMM(y, ones((n, 1)), QS)
lmm.fit(verbose=False)
plmm = LMMPredict(y, lmm.beta, lmm.v0, lmm.v1, lmm.mean(),
lmm.covariance())
K = dot(X, X.T)
pm = plmm.predictive_mean(ones((n, 1)), K, K.diagonal())
assert_allclose(corrcoef(y, pm)[0, 1], 0.8358820971891354)