def test_prior_knowledge_invalid():
# causal direction: x0 --> x1 --> x3
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
# prior knowledge: invalid
pk = np.array([
[0, -1, -1],
[-1, 0, -1],
[-1, -1, 0],
[-1, -1, -1],
])
try:
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
except ValueError:
pass
else:
raise AssertionError
def test_fit_success():
# causal direction: x0 --> x1 --> x3
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
# check the causal ordering
co = model.causal_order_
assert co.index(0) < co.index(1) < co.index(3)
# check the adjacency matrix
am = model.adjacency_matrices_[0]
assert am[1, 0] > 1.5 and am[3, 1] > 3.5
# check the adjacency matrix
am = model.adjacency_matrices_[1]
assert am[1, 0] > 1.5 and am[3, 1] > 3.5
am[1, 0] = 0.0
am[3, 1] = 0.0
assert np.sum(am) < 0.1
co = model.causal_order_
am = model.adjacency_matrices_
p_values = model.get_error_independence_p_values(X_list)
def test_bootstrap_invalid_data():
# causal direction: x0 --> x1 --> x3
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
# Invalid argument: bootstrap(n_sampling=-1)
model = MultiGroupDirectLiNGAM()
try:
result = model.bootstrap(X_list, n_sampling=-1)
except ValueError:
pass
else:
raise AssertionError
# Invalid argument: bootstrap(n_sampling='3')
model = MultiGroupDirectLiNGAM()
try:
result = model.bootstrap(X_list, n_sampling='3')
except ValueError:
pass
else:
raise AssertionError
def test_estimate_total_effect_invalid():
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
model._causal_order = [0, 1, 2, 3]
# warning
model.estimate_total_effect(X_list, 3, 1)
def test_fit_invalid_data():
# Different features
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2]).T, columns=['x0', 'x1', 'x2'])
X_list = [X1, X2]
try:
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
except ValueError:
pass
else:
raise AssertionError
# Not list data
X = 1
try:
model = MultiGroupDirectLiNGAM()
model.fit(X)
except ValueError:
pass
else:
raise AssertionError
# Include not-array data
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, 1]
try:
model = MultiGroupDirectLiNGAM()
model.fit(X)
except ValueError:
pass
else:
raise AssertionError
# Include non-numeric data
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=5)
x1 = np.array(['X', 'Y', 'X', 'Y', 'X'])
X2 = pd.DataFrame(np.array([x0, x1]).T, columns=['x0', 'x1'])
X_list = [X1, X2]
try:
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
except ValueError:
pass
else:
raise AssertionError
# Include NaN values
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1]).T, columns=['x0', 'x1'])
X2.iloc[100, 0] = np.nan
X_list = [X1, X2]
try:
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
except ValueError:
pass
else:
raise AssertionError
# Include infinite values
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
X = pd.DataFrame(np.array([x0, x1]).T, columns=['x0', 'x1'])
X.iloc[100, 0] = np.inf
X_list = [X1, X2]
try:
model = MultiGroupDirectLiNGAM()
model.fit(X_list)
except ValueError:
pass
else:
raise AssertionError
def test_prior_knowledge_success():
# causal direction: x0 --> x1 --> x3
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
# prior knowledge: nothing
pk = np.array([
[0, -1, -1, -1],
[-1, 0, -1, -1],
[-1, -1, 0, -1],
[-1, -1, -1, 0],
])
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
co = model.causal_order_
assert co.index(0) < co.index(1) < co.index(3)
# prior knowledge: x1 is exogenous
pk = np.array([
[0, -1, -1, -1],
[0, 0, 0, 0],
[-1, -1, 0, -1],
[-1, -1, -1, 0],
])
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
co = model.causal_order_
print(co)
assert co.index(1) < co.index(0)
assert co.index(1) < co.index(2)
assert co.index(1) < co.index(3)
# prior knowledge: x0 is sink
pk = np.array([
[0, -1, -1, -1],
[0, 0, -1, -1],
[0, -1, 0, -1],
[0, -1, -1, 0],
])
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
co = model.causal_order_
print(co)
assert co.index(0) > co.index(1)
assert co.index(0) > co.index(2)
assert co.index(0) > co.index(3)
# prior knowledge: x2-->x3 has path
pk = np.array([
[0, -1, -1, -1],
[-1, 0, -1, -1],
[-1, -1, 0, 0],
[-1, -1, 1, 0],
])
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
co = model.causal_order_
print(co)
assert co.index(2) < co.index(3)
# prior knowledge: x1-->x3 does not have path
pk = np.array([
[0, -1, -1, -1],
[-1, 0, -1, -1],
[-1, -1, 0, -1],
[-1, 0, -1, 0],
])
model = MultiGroupDirectLiNGAM(prior_knowledge=pk)
model.fit(X_list)
def test_bootstrap_success():
# causal direction: x0 --> x1 --> x3
x0 = np.random.uniform(size=1000)
x1 = 2.0 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.0 * x1 + np.random.uniform(size=1000)
X1 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
x0 = np.random.uniform(size=1000)
x1 = 2.1 * x0 + np.random.uniform(size=1000)
x2 = np.random.uniform(size=1000)
x3 = 4.2 * x1 + np.random.uniform(size=1000)
X2 = pd.DataFrame(np.array([x0, x1, x2, x3]).T,
columns=['x0', 'x1', 'x2', 'x3'])
X_list = [X1, X2]
model = MultiGroupDirectLiNGAM()
result_list = model.bootstrap(X_list, n_sampling=5)
result = result_list[0]
result.adjacency_matrices_
result.total_effects_
# No argument
cdc = result.get_causal_direction_counts()
# n_directions=2
cdc = result.get_causal_direction_counts(n_directions=2)
# min_causal_effect=0.2
cdc = result.get_causal_direction_counts(min_causal_effect=0.2)
# split_by_causal_effect_sign=True
cdc = result.get_causal_direction_counts(split_by_causal_effect_sign=True)
# No argument
dagc = result.get_directed_acyclic_graph_counts()
# n_dags=2
dagc = result.get_directed_acyclic_graph_counts(n_dags=2)
# min_causal_effect=0.6
dagc = result.get_directed_acyclic_graph_counts(min_causal_effect=0.6)
# split_by_causal_effect_sign=True
dagc = result.get_directed_acyclic_graph_counts(
split_by_causal_effect_sign=True)
# get_probabilities
probs = result.get_probabilities()
# get_probabilities
probs = result.get_probabilities(min_causal_effect=0.6)
# get_total_causal_effects
ce = result.get_total_causal_effects()
# get_total_causal_effects
ce = result.get_total_causal_effects(min_causal_effect=0.6)