def test_glmmexpfam_wrong_qs():
random = RandomState(0)
X = random.randn(10, 15)
linear_eye_cov().value()
QS = [0, 1]
ntri = random.randint(1, 30, 10)
nsuc = [random.randint(0, i) for i in ntri]
with pytest.raises(ValueError):
GLMMExpFam((nsuc, ntri), "binomial", X, QS)
def test_glmmnormal():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
M = random.randn(nsamples, 3)
K = linear_eye_cov().value()
QS = economic_qs(K)
eta = random.randn(nsamples)
tau = 10 * random.rand(nsamples)
glmm = GLMMNormal(eta, tau, X, QS)
glmm.beta = asarray([1.0, 0, 0.5, 0.1, 0.4])
assert_allclose(glmm.lml(), -19.284378946701814)
assert_allclose(glmm._check_grad(), 0, atol=1e-3, rtol=RTOL)
flmm = glmm.get_fast_scanner()
r = flmm.fast_scan(M, verbose=False)
assert_allclose(r["lml"], [9.64605678059, 9.17041834, 9.56927990771])
assert_allclose(r["effsizes1"],
[-0.0758297759308, 0.0509863368859, 0.0876858800519])
assert_allclose(
r["scale"],
[0.0053192483818597395, 0.005850105527002988, 0.00540155776161286])
def test_glmmexpfam_precise():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
QS = economic_qs(K)
ntri = random.randint(1, 30, nsamples)
nsuc = [random.randint(0, i) for i in ntri]
glmm = GLMMExpFam(nsuc, ["binomial", ntri], X, QS)
glmm.beta = asarray([1.0, 0, 0.5, 0.1, 0.4])
glmm.scale = 1.0
assert_allclose(glmm.lml(), -44.74191041468836, atol=ATOL, rtol=RTOL)
glmm.scale = 2.0
assert_allclose(glmm.lml(), -36.19907331929086, atol=ATOL, rtol=RTOL)
glmm.scale = 3.0
assert_allclose(glmm.lml(), -33.02139830387104, atol=ATOL, rtol=RTOL)
glmm.scale = 4.0
assert_allclose(glmm.lml(), -31.42553401678996, atol=ATOL, rtol=RTOL)
glmm.scale = 5.0
assert_allclose(glmm.lml(), -30.507029479473243, atol=ATOL, rtol=RTOL)
glmm.scale = 6.0
assert_allclose(glmm.lml(), -29.937569702301232, atol=ATOL, rtol=RTOL)
glmm.delta = 0.1
assert_allclose(glmm.lml(), -30.09977907145003, atol=ATOL, rtol=RTOL)
assert_allclose(glmm._check_grad(), 0, atol=1e-3, rtol=RTOL)
def test_glmmexpfam_qs_none():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
z = random.multivariate_normal(0.2 * ones(nsamples), K)
ntri = random.randint(1, 30, nsamples)
nsuc = zeros(nsamples, dtype=int)
for (i, ni) in enumerate(ntri):
nsuc[i] += sum(z[i] + 0.2 * random.randn(ni) > 0)
ntri = ascontiguousarray(ntri)
glmm = GLMMExpFam(nsuc, ("binomial", ntri), X, None)
assert_allclose(glmm.lml(), -38.30173374439622, atol=ATOL, rtol=RTOL)
glmm.fix("beta")
glmm.fix("scale")
glmm.fit(verbose=False)
assert_allclose(glmm.lml(), -32.03927471370041, atol=ATOL, rtol=RTOL)
glmm.unfix("beta")
glmm.unfix("scale")
glmm.fit(verbose=False)
assert_allclose(glmm.lml(), -19.575736561760586, atol=ATOL, rtol=RTOL)
def test_glmmexpfam_optimize():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
z = random.multivariate_normal(0.2 * ones(nsamples), K)
QS = economic_qs(K)
ntri = random.randint(1, 30, nsamples)
nsuc = zeros(nsamples, dtype=int)
for (i, ni) in enumerate(ntri):
nsuc[i] += sum(z[i] + 0.2 * random.randn(ni) > 0)
ntri = ascontiguousarray(ntri)
glmm = GLMMExpFam(nsuc, ("binomial", ntri), X, QS)
assert_allclose(glmm.lml(), -29.102168129099287, atol=ATOL, rtol=RTOL)
glmm.fix("beta")
glmm.fix("scale")
glmm.fit(verbose=False)
assert_allclose(glmm.lml(), -27.635788105778012, atol=ATOL, rtol=RTOL)
glmm.unfix("beta")
glmm.unfix("scale")
glmm.fit(verbose=False)
assert_allclose(glmm.lml(), -19.68486269551159, atol=ATOL, rtol=RTOL)
def test_glmmexpfam_scale_very_high():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
QS = economic_qs(K)
ntri = random.randint(1, 30, nsamples)
nsuc = [random.randint(0, i) for i in ntri]
glmm = GLMMExpFam(nsuc, ("binomial", ntri), X, QS)
glmm.beta = asarray([1.0, 0, 0.5, 0.1, 0.4])
glmm.scale = 30.0
assert_allclose(glmm.lml(), -29.632791380478736, atol=ATOL, rtol=RTOL)
assert_allclose(glmm._check_grad(), 0, atol=1e-3)
def test_glmmexpfam_delta1():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
QS = economic_qs(K)
ntri = random.randint(1, 30, nsamples)
nsuc = [random.randint(0, i) for i in ntri]
glmm = GLMMExpFam(nsuc, ("binomial", ntri), X, QS)
glmm.beta = asarray([1.0, 0, 0.5, 0.1, 0.4])
glmm.delta = 1
assert_allclose(glmm.lml(), -47.09677870648636, atol=ATOL, rtol=RTOL)
assert_allclose(glmm._check_grad(), 0, atol=1e-4)
def test_glmmexpfam_copy():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
z = random.multivariate_normal(0.2 * ones(nsamples), K)
QS = economic_qs(K)
ntri = random.randint(1, 30, nsamples)
nsuc = zeros(nsamples, dtype=int)
for (i, ni) in enumerate(ntri):
nsuc[i] += sum(z[i] + 0.2 * random.randn(ni) > 0)
ntri = ascontiguousarray(ntri)
glmm0 = GLMMExpFam(nsuc, ("binomial", ntri), X, QS)
assert_allclose(glmm0.lml(), -29.10216812909928, atol=ATOL, rtol=RTOL)
glmm0.fit(verbose=False)
v = -19.575736562427252
assert_allclose(glmm0.lml(), v)
glmm1 = glmm0.copy()
assert_allclose(glmm1.lml(), v)
glmm1.scale = 0.92
assert_allclose(glmm0.lml(), v, atol=ATOL, rtol=RTOL)
assert_allclose(glmm1.lml(), -30.832831740038056, atol=ATOL, rtol=RTOL)
glmm0.fit(verbose=False)
glmm1.fit(verbose=False)
v = -19.575736562378573
assert_allclose(glmm0.lml(), v)
assert_allclose(glmm1.lml(), v)
def test_glmmnormal_copy():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
QS = economic_qs(linear_eye_cov().value())
eta = random.randn(nsamples)
tau = random.rand(nsamples) * 10
glmm0 = GLMMNormal(eta, tau, X, QS)
assert_allclose(glmm0.lml(), -12.646439806030257, atol=ATOL, rtol=RTOL)
glmm0.fit(verbose=False)
v = -4.758450057194982
assert_allclose(glmm0.lml(), v, atol=ATOL, rtol=RTOL)
glmm1 = glmm0.copy()
assert_allclose(glmm1.lml(), v, atol=ATOL, rtol=RTOL)
glmm1.scale = 0.92
assert_allclose(glmm0.lml(), v, atol=ATOL, rtol=RTOL)
assert_allclose(glmm1.lml(), -10.986014936977927, atol=ATOL, rtol=RTOL)
glmm0.fit(verbose=False)
glmm1.fit(verbose=False)
assert_allclose(glmm0.lml(), v, atol=ATOL, rtol=RTOL)
assert_allclose(glmm1.lml(), v, atol=ATOL, rtol=RTOL)
K = asarray([
[
1.00000001e-03,
-2.36582704e-12,
-1.10745946e-12,
-5.95414742e-12,
1.32859621e-12,
-1.97576469e-12,
3.10002744e-12,
-2.98276215e-12,
2.21766464e-12,
-4.11149765e-13,
],
[
-2.36582704e-12,
1.00000000e-03,
3.38869300e-12,
-1.21382132e-12,
-1.18303301e-12,
2.99685550e-12,
-1.12667797e-12,
1.31258715e-14,
-3.41562159e-12,
-1.79390668e-12,
],
[
-1.10745946e-12,
3.38869300e-12,
1.00000001e-03,
-4.35318561e-12,
-2.94696950e-12,
3.70379702e-12,
1.15190236e-13,
-4.09334846e-12,
-6.56245537e-12,
-9.57373359e-13,
],
[
-5.95414742e-12,
-1.21382132e-12,
-4.35318561e-12,
1.00000002e-03,
-9.18763856e-13,
-4.95358110e-12,
-2.44724110e-12,
-1.70406838e-12,
1.15381873e-13,
3.10086508e-12,
],
[
1.32859621e-12,
-1.18303301e-12,
-2.94696950e-12,
-9.18763856e-13,
1.00000001e-03,
-1.55985955e-12,
-2.09090626e-12,
5.35213097e-13,
-2.05846838e-12,
-7.77741240e-13,
],
[
-1.97576469e-12,
2.99685550e-12,
3.70379702e-12,
-4.95358110e-12,
-1.55985955e-12,
1.00000001e-03,
2.16017366e-12,
-1.25544702e-12,
-3.18748695e-12,
-6.45778208e-12,
],
[
3.10002744e-12,
-1.12667797e-12,
1.15190236e-13,
-2.44724110e-12,
-2.09090626e-12,
2.16017366e-12,
1.00000001e-03,
-4.07492827e-12,
-2.49548897e-12,
-5.02704611e-12,
],
[
-2.98276215e-12,
1.31258715e-14,
-4.09334846e-12,
-1.70406838e-12,
5.35213097e-13,
-1.25544702e-12,
-4.07492827e-12,
1.00000002e-03,
-2.16846857e-12,
-1.08364499e-12,
],
[
2.21766464e-12,
-3.41562159e-12,
-6.56245537e-12,
1.15381873e-13,
-2.05846838e-12,
-3.18748695e-12,
-2.49548897e-12,
-2.16846857e-12,
1.00000001e-03,
4.83828810e-12,
],
[
-4.11149765e-13,
-1.79390668e-12,
-9.57373359e-13,
3.10086508e-12,
-7.77741240e-13,
-6.45778208e-12,
-5.02704611e-12,
-1.08364499e-12,
4.83828810e-12,
1.00000001e-03,
],
])
assert_allclose(glmm0.covariance(), K, rtol=1e-5, atol=1e-5)
assert_allclose(glmm1.covariance(), K, rtol=1e-5, atol=1e-5)
def test_glmmexpfam_glmmnormal_get_fast_scanner():
nsamples = 10
random = RandomState(0)
X = random.randn(nsamples, 5)
K = linear_eye_cov().value()
QS = economic_qs(K)
eta = random.randn(nsamples)
tau = 10 * random.rand(nsamples)
glmm = GLMMNormal(eta, tau, X, QS)
glmm.fit(verbose=False)
want = [-0.08228058, -0.03910674, 0.04226152, -0.05893827, 0.01718722]
assert_allclose(glmm.beta, want, atol=1e-3, rtol=1e-3)
assert_allclose(0.001, glmm.scale, atol=1e-3, rtol=1e-3)
assert_allclose(0.999999994119, glmm.delta, atol=1e-3, rtol=1e-3)
scanner = glmm.get_fast_scanner()
r = scanner.fast_scan(X, verbose=False)
assert_allclose(
r["lml"],
[
3.666664259270515,
3.6666642592705188,
3.666664259270515,
3.666664259270515,
3.6666642592705188,
],
rtol=1e-6,
)
assert_allclose(
r["effsizes0"],
[
array([
-0.04114011, -0.03910684, 0.04226118, -0.05893773, 0.01718666
]),
array([
-0.08228023, -0.01955342, 0.04226118, -0.05893773, 0.01718666
]),
array([
-0.08228023, -0.03910684, 0.02113059, -0.05893773, 0.01718666
]),
array([
-0.08228023, -0.03910684, 0.04226118, -0.02946886, 0.01718666
]),
array([
-0.08228023, -0.03910684, 0.04226118, -0.05893773, 0.00859333
]),
],
rtol=1e-6,
)
assert_allclose(
r["effsizes1"],
[
-0.04114011396191331,
-0.01955341813308458,
0.02113058926220273,
-0.029468864872716646,
0.008593328546792367,
],
rtol=1e-6,
)
assert_allclose(
r["scale"],
[
0.07341651661479422,
0.07341651661479419,
0.07341651661479422,
0.07341651661479422,
0.0734165166147942,
],
rtol=1e-6,
)
