def test_str_repr(self):
spa = SPA(self.benchmark, self.models)
expected = 'SPA(studentization: asymptotic, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
expected = expected[:-1] + ', ID: ' + hex(id(spa)) + ')'
assert_equal(spa.__repr__(), expected)
expected = ('<strong>SPA</strong>(' +
'<strong>studentization</strong>: asymptotic, ' +
'<strong>bootstrap</strong>: ' + str(spa.bootstrap) +
', <strong>ID</strong>: ' + hex(id(spa)) + ')')
assert_equal(spa._repr_html_(), expected)
spa = SPA(self.benchmark, self.models, studentize=False,
bootstrap='cbb')
expected = 'SPA(studentization: none, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
spa = SPA(self.benchmark, self.models, nested=True,
bootstrap='moving_block')
expected = 'SPA(studentization: bootstrap, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
def test_equivalence(self):
spa = SPA(self.benchmark, self.models, block_size=10, reps=100)
spa.seed(23456)
spa.compute()
numpy_pvalues = spa.pvalues
spa = SPA(self.benchmark_df, self.models_df, block_size=10, reps=100)
spa.seed(23456)
spa.compute()
pandas_pvalues = spa.pvalues
assert_series_equal(numpy_pvalues, pandas_pvalues)
def test_bootstrap_selection(self):
spa = SPA(self.benchmark, self.models, bootstrap='sb')
assert isinstance(spa.bootstrap, StationaryBootstrap)
spa = SPA(self.benchmark, self.models, bootstrap='cbb')
assert isinstance(spa.bootstrap, CircularBlockBootstrap)
spa = SPA(self.benchmark, self.models, bootstrap='circular')
assert isinstance(spa.bootstrap, CircularBlockBootstrap)
spa = SPA(self.benchmark, self.models, bootstrap='mbb')
assert isinstance(spa.bootstrap, MovingBlockBootstrap)
spa = SPA(self.benchmark, self.models, bootstrap='moving block')
assert isinstance(spa.bootstrap, MovingBlockBootstrap)
def test_bootstrap_selection(spa_data):
spa = SPA(spa_data.benchmark, spa_data.models, bootstrap="sb")
assert isinstance(spa.bootstrap, StationaryBootstrap)
spa = SPA(spa_data.benchmark, spa_data.models, bootstrap="cbb")
assert isinstance(spa.bootstrap, CircularBlockBootstrap)
spa = SPA(spa_data.benchmark, spa_data.models, bootstrap="circular")
assert isinstance(spa.bootstrap, CircularBlockBootstrap)
spa = SPA(spa_data.benchmark, spa_data.models, bootstrap="mbb")
assert isinstance(spa.bootstrap, MovingBlockBootstrap)
spa = SPA(spa_data.benchmark, spa_data.models, bootstrap="moving block")
assert isinstance(spa.bootstrap, MovingBlockBootstrap)
def test_str_repr(self):
spa = SPA(self.benchmark, self.models)
expected = 'SPA(studentization: asymptotic, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
expected = expected[:-1] + ', ID: ' + hex(id(spa)) + ')'
assert_equal(spa.__repr__(), expected)
expected = ('<strong>SPA</strong>(' +
'<strong>studentization</strong>: asymptotic, ' +
'<strong>bootstrap</strong>: ' + str(spa.bootstrap) +
', <strong>ID</strong>: ' + hex(id(spa)) + ')')
assert_equal(spa._repr_html_(), expected)
spa = SPA(self.benchmark,
self.models,
studentize=False,
bootstrap='cbb')
expected = 'SPA(studentization: none, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
spa = SPA(self.benchmark,
self.models,
nested=True,
bootstrap='moving_block')
expected = 'SPA(studentization: bootstrap, bootstrap: ' + \
str(spa.bootstrap) + ')'
assert_equal(str(spa), expected)
def test_str_repr(spa_data):
spa = SPA(spa_data.benchmark, spa_data.models)
expected = "SPA(studentization: asymptotic, bootstrap: " + str(
spa.bootstrap) + ")"
assert_equal(str(spa), expected)
expected = expected[:-1] + ", ID: " + hex(id(spa)) + ")"
assert_equal(spa.__repr__(), expected)
expected = ("<strong>SPA</strong>(" +
"<strong>studentization</strong>: asymptotic, " +
"<strong>bootstrap</strong>: " + str(spa.bootstrap) +
", <strong>ID</strong>: " + hex(id(spa)) + ")")
assert_equal(spa._repr_html_(), expected)
spa = SPA(spa_data.benchmark,
spa_data.models,
studentize=False,
bootstrap="cbb")
expected = "SPA(studentization: none, bootstrap: " + str(
spa.bootstrap) + ")"
assert_equal(str(spa), expected)
spa = SPA(spa_data.benchmark,
spa_data.models,
nested=True,
bootstrap="moving_block")
expected = "SPA(studentization: bootstrap, bootstrap: " + str(
spa.bootstrap) + ")"
assert_equal(str(spa), expected)
def test_seed_reset(self):
spa = SPA(self.benchmark, self.models, reps=10)
spa.seed(23456)
initial_state = spa.bootstrap.random_state
assert_equal(spa.bootstrap._seed, 23456)
spa.compute()
spa.reset()
assert_equal(spa._pvalues, None)
assert_equal(spa.bootstrap.random_state, initial_state)
def test_seed_reset(self):
spa = SPA(self.benchmark, self.models, reps=10)
spa.seed(23456)
initial_state = spa.bootstrap.random_state
assert_equal(spa.bootstrap._seed, 23456)
spa.compute()
spa.reset()
assert_equal(spa._pvalues, None)
assert_equal(spa.bootstrap.random_state, initial_state)
def test_superior_models(self):
adj_models = self.models - linspace(-0.4, 0.4, self.k)
stepm = StepM(self.benchmark, adj_models, reps=120)
stepm.compute()
superior_models = stepm.superior_models
assert len(superior_models) > 0
spa = SPA(self.benchmark, adj_models, reps=120)
spa.compute()
spa.pvalues
spa.critical_values(0.05)
spa.better_models(0.05)
adj_models = self.models_df - linspace(-3.0, 3.0, self.k)
stepm = StepM(self.benchmark_series, adj_models, reps=120)
stepm.compute()
superior_models = stepm.superior_models
assert len(superior_models) > 0
def test_pvalues_and_critvals(spa_data):
spa = SPA(spa_data.benchmark, spa_data.models, reps=100)
spa.compute()
spa.seed(23456)
simulated_vals = spa._simulated_vals
max_stats = np.max(simulated_vals, 0)
max_loss_diff = np.max(spa._loss_diff.mean(0), 0)
pvalues = np.mean(max_loss_diff <= max_stats, 0)
pvalues = pd.Series(pvalues, index=["lower", "consistent", "upper"])
assert_series_equal(pvalues, spa.pvalues)
crit_vals = np.percentile(max_stats, 90.0, axis=0)
crit_vals = pd.Series(crit_vals, index=["lower", "consistent", "upper"])
assert_series_equal(spa.critical_values(0.10), crit_vals)
def test_superior_models(self):
adj_models = self.models - linspace(-1.0, 1.0, self.k)
stepm = StepM(self.benchmark, adj_models, reps=120)
stepm.compute()
superior_models = stepm.superior_models
assert len(superior_models) > 0
spa = SPA(self.benchmark, adj_models, reps=120)
spa.compute()
spa.pvalues
spa.critical_values(0.05)
spa.better_models(0.05)
adj_models = self.models_df - linspace(-3.0, 3.0, self.k)
stepm = StepM(self.benchmark_series, adj_models, reps=120)
stepm.compute()
superior_models = stepm.superior_models
assert len(superior_models) > 0
def test_pvalues_and_critvals(self):
spa = SPA(self.benchmark, self.models, reps=100)
spa.compute()
spa.seed(23456)
simulated_vals = spa._simulated_vals
max_stats = np.max(simulated_vals, 0)
max_loss_diff = np.max(spa._loss_diff.mean(0), 0)
pvalues = np.mean(max_loss_diff <= max_stats, 0)
pvalues = pd.Series(pvalues, index=['lower', 'consistent', 'upper'])
assert_series_equal(pvalues, spa.pvalues)
crit_vals = np.percentile(max_stats, 90.0, axis=0)
crit_vals = pd.Series(crit_vals,
index=['lower', 'consistent', 'upper'])
assert_series_equal(spa.critical_values(0.10), crit_vals)
def test_variances_and_selection(self):
adj_models = self.models + linspace(-2, 0.5, self.k)
spa = SPA(self.benchmark, adj_models, block_size=10, reps=10)
spa.seed(23456)
spa.compute()
variances = spa._loss_diff_var
loss_diffs = spa._loss_diff
demeaned = spa._loss_diff - loss_diffs.mean(0)
t = loss_diffs.shape[0]
kernel_weights = np.zeros(t)
p = 1 / 10.0
for i in range(1, t):
kernel_weights[i] = ((1.0 - (i / t)) * ((1 - p) ** i)) + \
((i / t) * ((1 - p) ** (t - i)))
direct_vars = (demeaned ** 2).sum(0) / t
for i in range(1, t):
direct_vars += 2 * kernel_weights[i] * \
(demeaned[:t - i, :] * demeaned[i:, :]).sum(0) / t
assert_allclose(direct_vars, variances)
selection_criteria = -1.0 * np.sqrt((direct_vars / t) * 2 *
np.log(np.log(t)))
valid = loss_diffs.mean(0) >= selection_criteria
assert_equal(valid, spa._valid_columns)
# Bootstrap variances
spa = SPA(self.benchmark, self.models, block_size=10, reps=100,
nested=True)
spa.seed(23456)
spa.compute()
spa.reset()
bs = spa.bootstrap.clone(demeaned)
variances = spa._loss_diff_var
bootstrap_variances = t * bs.var(lambda x: x.mean(0), reps=100,
recenter=True)
assert_allclose(bootstrap_variances, variances)
def test_spa_nested(spa_data):
spa = SPA(spa_data.benchmark, spa_data.models, nested=True, reps=100)
spa.compute()
def test_single_model(self):
spa = SPA(self.benchmark, self.models[:, 0])
spa.compute()
spa = SPA(self.benchmark_series, self.models_df.iloc[:, 0])
spa.compute()
def test_spa_nested(self):
spa = SPA(self.benchmark, self.models, nested=True, reps=100)
spa.compute()
def test_single_model(spa_data):
spa = SPA(spa_data.benchmark, spa_data.models[:, 0])
spa.compute()
spa = SPA(spa_data.benchmark_series, spa_data.models_df.iloc[:, 0])
spa.compute()
def test_spa_nested(self):
spa = SPA(self.benchmark, self.models, nested=True, reps=100)
spa.compute()
def test_errors(self):
spa = SPA(self.benchmark, self.models, reps=100)
with pytest.raises(RuntimeError):
spa.pvalues
with pytest.raises(RuntimeError):
spa.critical_values()
with pytest.raises(RuntimeError):
spa.better_models()
with pytest.raises(ValueError):
SPA(self.benchmark, self.models, bootstrap='unknown')
spa.compute()
with pytest.raises(ValueError):
spa.better_models(pvalue_type='unknown')
with pytest.raises(ValueError):
spa.critical_values(pvalue=1.0)
def test_variances_and_selection(self):
adj_models = self.models + linspace(-2, 0.5, self.k)
spa = SPA(self.benchmark, adj_models, block_size=10, reps=10)
spa.seed(23456)
spa.compute()
variances = spa._loss_diff_var
loss_diffs = spa._loss_diff
demeaned = spa._loss_diff - loss_diffs.mean(0)
t = loss_diffs.shape[0]
kernel_weights = np.zeros(t)
p = 1 / 10.0
for i in range(1, t):
kernel_weights[i] = ((1.0 - (i / t)) * ((1 - p) ** i)) + \
((i / t) * ((1 - p) ** (t - i)))
direct_vars = (demeaned**2).sum(0) / t
for i in range(1, t):
direct_vars += 2 * kernel_weights[i] * \
(demeaned[:t - i, :] * demeaned[i:, :]).sum(0) / t
assert_allclose(direct_vars, variances)
selection_criteria = -1.0 * np.sqrt(
(direct_vars / t) * 2 * np.log(np.log(t)))
valid = loss_diffs.mean(0) >= selection_criteria
assert_equal(valid, spa._valid_columns)
# Bootstrap variances
spa = SPA(self.benchmark,
self.models,
block_size=10,
reps=100,
nested=True)
spa.seed(23456)
spa.compute()
spa.reset()
bs = spa.bootstrap.clone(demeaned)
variances = spa._loss_diff_var
bootstrap_variances = t * bs.var(
lambda x: x.mean(0), reps=100, recenter=True)
assert_allclose(bootstrap_variances, variances)
def test_single_model(self):
spa = SPA(self.benchmark, self.models[:, 0])
spa.compute()
spa = SPA(self.benchmark_series, self.models_df.iloc[:, 0])
spa.compute()
def test_errors(self):
spa = SPA(self.benchmark, self.models, reps=100)
with pytest.raises(RuntimeError):
spa.pvalues
with pytest.raises(RuntimeError):
spa.critical_values()
with pytest.raises(RuntimeError):
spa.better_models()
with pytest.raises(ValueError):
SPA(self.benchmark, self.models, bootstrap='unknown')
spa.compute()
with pytest.raises(ValueError):
spa.better_models(pvalue_type='unknown')
with pytest.raises(ValueError):
spa.critical_values(pvalue=1.0)
def test_equivalence(self):
spa = SPA(self.benchmark, self.models, block_size=10, reps=100)
spa.seed(23456)
spa.compute()
numpy_pvalues = spa.pvalues
spa = SPA(self.benchmark_df, self.models_df, block_size=10, reps=100)
spa.seed(23456)
spa.compute()
pandas_pvalues = spa.pvalues
assert_series_equal(numpy_pvalues, pandas_pvalues)