def test_predict(self):
model = VAR(2)
model.fit(data)
forecast, ci95_low, ci95_high = model.predict(10)
self.assertTrue(
np.logical_and(forecast > ci95_low, forecast < ci95_high).all())
def test_predict_as_table(self):
model = VAR(2)
model.fit(data)
forecast = model.predict(10, as_table=True)
self.assertEqual(len(forecast.domain.variables), 2 * (1 + 2))
super().closeEvent(event)
def closeContext(self) -> None:
"""
Gather configs in contextVariables and close context.
"""
if not self.features:
# only close in case of when features are not present if they are
# feature selection is defined by the input and context should
# not have impact
attrs, is_logit = [], []
for config in self.configs:
attrs.append(config.get_selection())
is_logit.append(config.is_logarithmic)
self.attrs = attrs
self.is_logit = is_logit
super().closeContext()
if __name__ == "__main__":
from orangecontrib.timeseries import ARIMA, VAR
airpassengers = Timeseries.from_file('airpassengers')
msft = airpassengers.interp()
model1 = ARIMA((3, 1, 1)).fit(airpassengers)
model2 = VAR(4).fit(msft)
ow = WidgetPreview(OWLineChart)
ow.run(set_data=airpassengers,
set_forecast=[(model1.predict(10, as_table=True), 0),
(model2.predict(10, as_table=True), 1)])