def test_summary(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(update_freq=0, disp=DISPLAY)
res.summary()
am = arch_model(self.y, mean='ar', lags=[1, 3, 5], dist='studentst')
res = am.fit(update_freq=0, disp=DISPLAY)
res.summary()
def test_1d_exog():
y = np.random.standard_normal((300))
x = np.random.standard_normal((300))
am = arch_model(y, x, mean="ARX", lags=2, vol="ARCH", q=0)
res = am.fit()
am = arch_model(y, x[:, None], mean="ARX", lags=2, vol="ARCH", q=0)
res2 = am.fit()
assert_series_equal(res.params, res2.params)
def test_starting_values(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(cov_type='mle', update_freq=0)
res2 = am.fit(starting_values=res.params, update_freq=0)
am = arch_model(self.y, mean='zero')
sv = np.array([1.0, 0.3, 0.8])
with warnings.catch_warnings(record=True) as w:
am.fit(starting_values=sv, update_freq=0)
assert_equal(len(w), 1)
def test_starting_values(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(cov_type='mle', update_freq=0)
res2 = am.fit(starting_values=res.params, update_freq=0)
am = arch_model(self.y, mean='zero')
sv = np.array([1.0, 0.3, 0.8])
with warnings.catch_warnings(record=True) as w:
am.fit(starting_values=sv, update_freq=0)
assert_equal(len(w), 1)
def test_optimization_options(self):
am = arch_model(None)
data = am.simulate([0.0, 0.1, 0.1, 0.85], 2500)
am = arch_model(data.data)
std = am.fit(disp='off')
loose = am.fit(tol=1e-2, disp='off')
assert std.loglikelihood != loose.loglikelihood
with warnings.catch_warnings(record=True) as w:
short = am.fit(options={'maxiter': 3}, disp='off')
assert len(w) == 1
assert std.loglikelihood != short.loglikelihood
assert short.convergence_flag != 0
def test_backcast_restricted(simulated_data):
# GH 440
mod = arch_model(simulated_data)
res = mod.fit(disp="off")
subset = (simulated_data[100:600] if isinstance(simulated_data, np.ndarray)
else simulated_data.iloc[100:600])
mod_restricted = arch_model(subset)
res_restricted = mod_restricted.fit(disp="off")
res_limited = mod.fit(first_obs=100, last_obs=600, disp="off")
assert_almost_equal(res_restricted.model._backcast,
res_restricted.model._backcast)
assert np.abs(res.model._backcast - res_limited.model._backcast) > 1e-8
def test_model_obs_equivalence(self):
"""Tests models that should use the same observation"""
am = arch_model(self.y_series.iloc[100:900])
res = am.fit(disp='off')
am = arch_model(self.y_series)
res2 = am.fit(disp='off', first_obs=100, last_obs=900)
index = self.y_series.index
res3 = am.fit(disp='off', first_obs=index[100], last_obs=index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res2.params.values, res3.params.values)
am = arch_model(self.y_series, hold_back=100)
res4 = am.fit(disp='off', last_obs=900)
assert_equal(res.params.values, res4.params.values)
def test_model_obs_equivalence(self):
"""Tests models that should use the same observation"""
am = arch_model(self.y_series.iloc[100:900])
res = am.fit(disp='off')
am = arch_model(self.y_series)
res2 = am.fit(disp='off', first_obs=100, last_obs=900)
index = self.y_series.index
res3 = am.fit(disp='off', first_obs=index[100], last_obs=index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res2.params.values, res3.params.values)
am = arch_model(self.y_series, hold_back=100)
res4 = am.fit(disp='off', last_obs=900)
assert_equal(res.params.values, res4.params.values)
def test_starting_values(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(cov_type='mle', update_freq=0, disp=DISPLAY)
res2 = am.fit(starting_values=res.params, update_freq=0, disp=DISPLAY)
assert isinstance(res, ARCHModelResult)
assert isinstance(res2, ARCHModelResult)
assert len(res.params) == 7
assert len(res2.params) == 7
am = arch_model(self.y, mean='zero')
sv = np.array([1.0, 0.3, 0.8])
with warnings.catch_warnings(record=True) as w:
am.fit(starting_values=sv, update_freq=0, disp=DISPLAY)
assert_equal(len(w), 1)
def test_starting_values(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(cov_type='mle', update_freq=0, disp=DISPLAY)
res2 = am.fit(starting_values=res.params, update_freq=0, disp=DISPLAY)
assert isinstance(res, ARCHModelResult)
assert isinstance(res2, ARCHModelResult)
assert len(res.params) == 7
assert len(res2.params) == 7
am = arch_model(self.y, mean='zero')
sv = np.array([1.0, 0.3, 0.8])
with warnings.catch_warnings(record=True) as w:
am.fit(starting_values=sv, update_freq=0, disp=DISPLAY)
assert_equal(len(w), 1)
def test_optimization_options(self):
norm = Normal(random_state=RandomState([12891298, 843084]))
am = arch_model(None)
am.distribution = norm
data = am.simulate(np.array([0.0, 0.1, 0.1, 0.85]), 2500)
am = arch_model(data.data)
std = am.fit(disp=DISPLAY)
loose = am.fit(tol=1e-2, disp=DISPLAY)
assert std.loglikelihood >= loose.loglikelihood
with warnings.catch_warnings(record=True) as w:
short = am.fit(options={"maxiter": 3}, disp=DISPLAY)
assert len(w) == 1
assert std.loglikelihood >= short.loglikelihood
assert short.convergence_flag != 0
def test_model_obs_equivalence_ar(self):
"""Tests models that should use the same observation"""
am = arch_model(self.y_series.iloc[100:900], mean="AR", lags=[1, 2, 4])
res = am.fit(disp=DISPLAY)
am = arch_model(self.y_series, mean="AR", lags=[1, 2, 4])
res2 = am.fit(disp=DISPLAY, first_obs=100, last_obs=900)
index = self.y_series.index
res3 = am.fit(disp=DISPLAY, first_obs=index[100], last_obs=index[900])
assert_almost_equal(res.params.values, res2.params.values, decimal=4)
assert_almost_equal(res2.params.values, res3.params.values, decimal=4)
am = arch_model(self.y_series, mean="AR", lags=[1, 2, 4], hold_back=100)
res4 = am.fit(disp=DISPLAY, first_obs=4, last_obs=900)
assert_almost_equal(res.params.values, res4.params.values, decimal=4)
assert am.hold_back == 100
def test_sample_adjustment(self):
am = arch_model(self.y_series, vol='Constant')
res = am.fit(disp='off')
res_adj = am.fit(disp='off',
first_obs=0,
last_obs=self.y_series.shape[0] + 1)
assert_equal(res.resid.values, res_adj.resid.values)
assert_equal(res.params.values, res_adj.params.values)
res = am.fit(disp='off', first_obs=100)
assert res.fit_start == 100
res_adj = am.fit(disp='off', first_obs=self.y_series.index[100])
assert_equal(res.params.values, res_adj.params.values)
assert_equal(res.resid.values, res_adj.resid.values)
res = am.fit(disp='off', last_obs=900)
res2 = am.fit(disp='off', last_obs=self.y_series.index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res.resid.values, res2.resid.values)
res = am.fit(disp='off', first_obs=100, last_obs=900)
res2 = am.fit(disp='off',
first_obs=self.y_series.index[100],
last_obs=self.y_series.index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res.resid.values, res2.resid.values)
def test_sample_adjustment(self):
am = arch_model(self.y_series, vol='Constant')
res = am.fit(disp='off')
res_adj = am.fit(disp='off',
first_obs=0,
last_obs=self.y_series.shape[0] + 1)
assert_equal(res.resid.values, res_adj.resid.values)
assert_equal(res.params.values, res_adj.params.values)
res = am.fit(disp='off', first_obs=100)
assert res.fit_start == 100
res_adj = am.fit(disp='off', first_obs=self.y_series.index[100])
assert_equal(res.params.values, res_adj.params.values)
assert_equal(res.resid.values, res_adj.resid.values)
res = am.fit(disp='off', last_obs=900)
res2 = am.fit(disp='off', last_obs=self.y_series.index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res.resid.values, res2.resid.values)
res = am.fit(disp='off', first_obs=100, last_obs=900)
res2 = am.fit(disp='off',
first_obs=self.y_series.index[100],
last_obs=self.y_series.index[900])
assert_equal(res.params.values, res2.params.values)
assert_equal(res.resid.values, res2.resid.values)
def test_model_obs_equivalence_ar(self):
"""Tests models that should use the same observation"""
am = arch_model(self.y_series.iloc[100:900], mean='AR', lags=[1, 2, 4])
res = am.fit(disp='off')
am = arch_model(self.y_series, mean='AR', lags=[1, 2, 4])
res2 = am.fit(disp='off', first_obs=100, last_obs=900)
index = self.y_series.index
res3 = am.fit(disp='off', first_obs=index[100], last_obs=index[900])
assert_almost_equal(res.params.values, res2.params.values)
assert_almost_equal(res2.params.values, res3.params.values)
am = arch_model(self.y_series, mean='AR', lags=[1, 2, 4],
hold_back=100)
res4 = am.fit(disp='off', first_obs=4, last_obs=900)
assert_almost_equal(res.params.values, res4.params.values, decimal=4)
assert am.hold_back == 100
def test_first_after_last(self):
am = arch_model(self.y_series)
with pytest.raises(ValueError):
am.fit(disp='off', first_obs=500, last_obs=480)
with pytest.raises(ValueError):
am.fit(disp='off',
first_obs=self.y_series.index[500],
last_obs=self.y_series.index[480])
def test_first_after_last(self):
am = arch_model(self.y_series)
with pytest.raises(ValueError):
am.fit(disp='off', first_obs=500, last_obs=480)
with pytest.raises(ValueError):
am.fit(disp='off',
first_obs=self.y_series.index[500],
last_obs=self.y_series.index[480])
def test_summary(self):
am = arch_model(self.y, mean='ar', lags=[1, 3, 5])
res = am.fit(update_freq=0, disp=DISPLAY)
res.summary()
am = arch_model(self.y, mean='ar', lags=[1, 3, 5], dist='studentst')
assert isinstance(am.distribution, StudentsT)
res = am.fit(update_freq=0, disp=DISPLAY)
res.summary()
am = arch_model(self.y, mean='ar', lags=[1, 3, 5], dist='ged')
assert isinstance(am.distribution, GeneralizedError)
res = am.fit(update_freq=0, disp=DISPLAY)
res.summary()
am = arch_model(self.y, mean='ar', lags=[1, 3, 5], dist='skewt')
res = am.fit(update_freq=0, disp=DISPLAY)
assert isinstance(am.distribution, SkewStudent)
res.summary()
def test_autoscale():
rs = np.random.RandomState(34254321)
dist = Normal(random_state=rs)
am = arch_model(None)
am.distribution = dist
data = am.simulate([0, 0.0001, 0.05, 0.94], nobs=1000)
am = arch_model(data.data)
with pytest.warns(DataScaleWarning):
res = am.fit(disp=DISPLAY)
assert_almost_equal(res.scale, 1.0)
am = arch_model(data.data, rescale=True)
res_auto = am.fit(disp=DISPLAY)
assert_almost_equal(res_auto.scale, 10.0)
am = arch_model(10 * data.data)
res_manual = am.fit(disp=DISPLAY)
assert_series_equal(res_auto.params, res_manual.params)
res_no = arch_model(data.data, rescale=False).fit(disp=DISPLAY)
assert res_no.scale == 1.0
am = arch_model(10000 * data.data, rescale=True)
res_big = am.fit(disp=DISPLAY)
assert_almost_equal(res_big.scale, 0.1)
def test_fixed_user_parameters(self):
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY)
fixed_res = am.fix(res.params)
assert_series_equal(res.conditional_volatility,
fixed_res.conditional_volatility)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
# Smoke for summary
fixed_res.summary()
def test_output_options(self):
am = arch_model(self.y_series)
orig_stdout = sys.stdout
try:
sio = StringIO()
sys.stdout = sio
am.fit(disp=DISPLAY)
sio.seek(0)
output = sio.read()
assert len(output) == 0
finally:
sys.stdout = orig_stdout
def test_fixed_user_parameters(self):
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY)
fixed_res = am.fix(res.params)
assert_series_equal(res.conditional_volatility,
fixed_res.conditional_volatility)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
# Smoke for summary
fixed_res.summary()
def test_fixed_user_parameters_new_model(self):
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY)
new_am = arch_model(self.y_series)
fixed_res = new_am.fix(res.params)
assert_series_equal(res.conditional_volatility,
fixed_res.conditional_volatility)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
# Test first and last dates
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY, first_obs=100, last_obs=900)
new_am = arch_model(self.y_series)
fixed_res = new_am.fix(res.params, first_obs=100, last_obs=900)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
def test_fixed_user_parameters_new_model(self):
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY)
new_am = arch_model(self.y_series)
fixed_res = new_am .fix(res.params)
assert_series_equal(res.conditional_volatility,
fixed_res.conditional_volatility)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
# Test first and last dates
am = arch_model(self.y_series)
res = am.fit(disp=DISPLAY, first_obs=100, last_obs=900)
new_am = arch_model(self.y_series)
fixed_res = new_am .fix(res.params, first_obs=100, last_obs=900)
assert_series_equal(res.params, fixed_res.params)
assert_equal(res.aic, fixed_res.aic)
assert_equal(res.bic, fixed_res.bic)
assert_equal(res.loglikelihood, fixed_res.loglikelihood)
assert_equal(res.num_params, fixed_res.num_params)
def test_output_options(self):
import sys
from arch.compat.python import StringIO
am = arch_model(self.y_series)
orig_stdout = sys.stdout
try:
sio = StringIO()
sys.stdout = sio
res = am.fit(disp='final')
sio.seek(0)
print('SIO!')
print(sio.read())
finally:
sys.stdout = orig_stdout
res = am.fit(disp='off')
def test_output_options(self):
import sys
from arch.compat.python import StringIO
am = arch_model(self.y_series)
orig_stdout = sys.stdout
try:
sio = StringIO()
sys.stdout = sio
res = am.fit(disp='final')
sio.seek(0)
print('SIO!')
print(sio.read())
finally:
sys.stdout = orig_stdout
res = am.fit(disp='off')
def test_convergence_warning(self):
y = np.array([0.83277114, 0.45194014, -0.33475561, -0.49463896,
0.54715787, 1.11895382, 1.31280266, 0.81464021,
0.8532107, 1.0967188, 0.9346354, 0.92289249, 1.01339085,
1.071065, 1.42413486, 1.15392453, 1.10929691, 0.96162061,
0.96489515, 0.93250153, 1.34509807, 1.80951607,
1.66313783, 1.38610821, 1.26381761])
am = arch_model(y, mean='ARX', lags=10, p=5, q=0)
with pytest.warns(ConvergenceWarning):
am.fit(disp=DISPLAY)
with pytest.warns(ConvergenceWarning):
am.fit(show_warning=True, disp=DISPLAY)
with pytest.warns(DataScaleWarning):
am.fit(show_warning=False, disp=DISPLAY)
def test_convergence_warning(self):
y = np.array([0.83277114, 0.45194014, -0.33475561, -0.49463896, 0.54715787,
1.11895382, 1.31280266, 0.81464021, 0.8532107, 1.0967188,
0.9346354, 0.92289249, 1.01339085, 1.071065, 1.42413486,
1.15392453, 1.10929691, 0.96162061, 0.96489515, 0.93250153,
1.34509807, 1.80951607, 1.66313783, 1.38610821, 1.26381761])
am = arch_model(y, mean='ARX', lags=10, p=5, q=0)
with warnings.catch_warnings(record=True) as w:
am.fit()
assert_equal(len(w), 1)
with warnings.catch_warnings(record=True) as w:
am.fit(show_warning=False)
assert_equal(len(w), 0)
with warnings.catch_warnings(record=True) as w:
am.fit(show_warning=True)
assert_equal(len(w), 1)
def test_missing_data_exception():
y = np.random.standard_normal(1000)
y[::29] = np.nan
with pytest.raises(ValueError, match="NaN or inf values"):
arch_model(y)
y = np.random.standard_normal(1000)
y[::53] = np.inf
with pytest.raises(ValueError, match="NaN or inf values"):
arch_model(y)
y[::29] = np.nan
y[::53] = np.inf
with pytest.raises(ValueError, match="NaN or inf values"):
arch_model(y)
def test_convergence_warning(self):
y = np.array([0.83277114, 0.45194014, -0.33475561, -0.49463896,
0.54715787, 1.11895382, 1.31280266, 0.81464021,
0.8532107, 1.0967188, 0.9346354, 0.92289249, 1.01339085,
1.071065, 1.42413486, 1.15392453, 1.10929691, 0.96162061,
0.96489515, 0.93250153, 1.34509807, 1.80951607,
1.66313783, 1.38610821, 1.26381761])
am = arch_model(y, mean='ARX', lags=10, p=5, q=0)
with warnings.catch_warnings(record=True) as w:
am.fit(disp=DISPLAY)
assert_equal(len(w), 1)
with warnings.catch_warnings(record=True) as w:
am.fit(show_warning=False, disp=DISPLAY)
assert_equal(len(w), 0)
with warnings.catch_warnings(record=True) as w:
am.fit(show_warning=True, disp=DISPLAY)
assert_equal(len(w), 1)
def test_output_options(self):
am = arch_model(self.y_series)
orig_stdout = sys.stdout
try:
sio = StringIO()
sys.stdout = sio
am.fit(disp='final')
sio.seek(0)
print('SIO!')
print(sio.read())
finally:
sys.stdout = orig_stdout
try:
sio = StringIO()
sys.stdout = sio
am.fit(disp='off')
sio.seek(0)
output = sio.read()
assert len(output) == 0
finally:
sys.stdout = orig_stdout
def test_arch_model(self):
am = arch_model(self.y)
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, GARCH)
assert isinstance(am.distribution, Normal)
am = arch_model(self.y, mean='harx', lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='har', lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, self.x, mean='ls')
assert isinstance(am, LS)
assert isinstance(am.volatility, GARCH)
am.__repr__()
am = arch_model(self.y, mean='arx', lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='ar', lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='ar', lags=None)
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='zero')
assert isinstance(am, ZeroMean)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, vol='Harch')
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, HARCH)
am = arch_model(self.y, vol='Constant')
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, ConstantVariance)
am = arch_model(self.y, vol='arch')
assert isinstance(am.volatility, ARCH)
am = arch_model(self.y, vol='egarch')
assert isinstance(am.volatility, EGARCH)
with pytest.raises(ValueError):
arch_model(self.y, mean='unknown')
with pytest.raises(ValueError):
arch_model(self.y, vol='unknown')
with pytest.raises(ValueError):
arch_model(self.y, dist='unknown')
am.fit(disp=DISPLAY)
def test_pandas(self):
am = arch_model(self.y_df, self.x_df, mean='ls')
assert isinstance(am, LS)
def test_invalid_arch_model():
with pytest.raises(TypeError, match="p must be"):
arch_model(SP500, p="3")
def test_no_variance():
mod = arch_model(np.ones(100))
with pytest.warns(ConvergenceWarning):
mod.fit(disp=DISPLAY)
def test_pandas(self):
am = arch_model(self.y_df, self.x_df, mean='ls')
assert isinstance(am, LS)
def test_arch_model(self):
am = arch_model(self.y)
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, GARCH)
assert isinstance(am.distribution, Normal)
am = arch_model(self.y, mean='harx', lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='har', lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, self.x, mean='ls')
assert isinstance(am, LS)
assert isinstance(am.volatility, GARCH)
am.__repr__()
am = arch_model(self.y, mean='arx', lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='ar', lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='ar', lags=None)
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean='zero')
assert isinstance(am, ZeroMean)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, vol='Harch')
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, HARCH)
am = arch_model(self.y, vol='Constant')
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, ConstantVariance)
am = arch_model(self.y, vol='arch')
assert isinstance(am.volatility, ARCH)
am = arch_model(self.y, vol='egarch')
assert isinstance(am.volatility, EGARCH)
with pytest.raises(ValueError):
arch_model(self.y, mean='unknown')
with pytest.raises(ValueError):
arch_model(self.y, vol='unknown')
with pytest.raises(ValueError):
arch_model(self.y, dist='unknown')
am.fit(disp=DISPLAY)
def test_arch_model(self):
am = arch_model(self.y)
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, GARCH)
assert isinstance(am.distribution, Normal)
am = arch_model(self.y, mean="harx", lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean="har", lags=[1, 5, 22])
assert isinstance(am, HARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, self.x, mean="ls")
assert isinstance(am, LS)
assert isinstance(am.volatility, GARCH)
am.__repr__()
am = arch_model(self.y, mean="arx", lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean="ar", lags=[1, 5, 22])
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean="ar", lags=None)
assert isinstance(am, ARX)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, mean="zero")
assert isinstance(am, ZeroMean)
assert isinstance(am.volatility, GARCH)
am = arch_model(self.y, vol="Harch")
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, HARCH)
am = arch_model(self.y, vol="Constant")
assert isinstance(am, ConstantMean)
assert isinstance(am.volatility, ConstantVariance)
am = arch_model(self.y, vol="arch")
assert isinstance(am.volatility, ARCH)
am = arch_model(self.y, vol="egarch")
assert isinstance(am.volatility, EGARCH)
am = arch_model(self.y, vol="figarch")
assert isinstance(am.volatility, FIGARCH)
with pytest.raises(ValueError):
arch_model(self.y, mean="unknown")
with pytest.raises(ValueError):
arch_model(self.y, vol="unknown")
with pytest.raises(ValueError):
arch_model(self.y, dist="unknown")
am.fit(disp=DISPLAY)
