def test_strategy_mean_seasonal(fh, sp, window_length):
if (window_length is not None
and window_length > sp) or (window_length is None):
f = NaiveForecaster(strategy="mean",
sp=sp,
window_length=window_length)
f.fit(y_train)
y_pred = f.predict(fh)
# check predicted index
assert_correct_pred_time_index(y_pred.index, y_train.index[-1], fh)
if window_length is None:
window_length = len(y_train)
# check values
fh = check_fh(fh) # get well formatted fh
reps = np.int(np.ceil(max(fh) / sp))
last_window = y_train.iloc[-window_length:].values
last_window = np.pad(
last_window,
(0, sp - len(last_window) % sp),
"constant",
constant_values=np.nan,
)
last_window = last_window.reshape(
np.int(np.ceil(len(last_window) / sp)), sp)
expected = np.tile(np.nanmean(last_window, axis=0), reps=reps)[fh - 1]
np.testing.assert_array_equal(y_pred, expected)
def test_predict_in_sample(Forecaster, fh, y_train):
f = _construct_instance(Forecaster)
try:
f.fit(y_train, fh=fh)
y_pred = f.predict()
assert_correct_pred_time_index(y_pred, y_train, fh)
except NotImplementedError:
pass
def test_predict_in_sample_full(Forecaster, y_train):
f = _construct_instance(Forecaster)
fh = -np.arange(len(y_train)) # full in-sample fh
try:
f.fit(y_train, fh=fh)
y_pred = f.predict()
assert_correct_pred_time_index(y_pred, y_train, fh)
except NotImplementedError:
pass
def test_predict_time_index(Forecaster, index_type, fh_type, is_relative,
steps):
y_train = make_forecasting_problem(index_type=index_type)
cutoff = y_train.index[-1]
fh = _make_fh(cutoff, steps, fh_type, is_relative)
f = _construct_instance(Forecaster)
try:
f.fit(y_train, fh=fh)
y_pred = f.predict()
assert_correct_pred_time_index(y_pred.index, y_train.index[-1], fh)
except NotImplementedError:
pass
def test_strategy_last_seasonal(fh, sp):
f = NaiveForecaster(strategy="last", sp=sp)
f.fit(y_train)
y_pred = f.predict(fh)
# check predicted index
assert_correct_pred_time_index(y_pred.index, y_train.index[-1], fh)
# check values
fh = check_fh(fh) # get well formatted fh
reps = np.int(np.ceil(max(fh) / sp))
expected = np.tile(y_train.iloc[-sp:], reps=reps)[fh - 1]
np.testing.assert_array_equal(y_pred, expected)
def check_pred_ints(pred_ints, y_train, y_pred, fh):
# make iterable
if isinstance(pred_ints, pd.DataFrame):
pred_ints = [pred_ints]
for pred_int in pred_ints:
assert list(pred_int.columns) == ["lower", "upper"]
assert_correct_pred_time_index(pred_int.index, y_train.index[-1], fh)
# check if errors are weakly monotonically increasing
pred_errors = y_pred - pred_int["lower"]
# assert pred_errors.is_mononotic_increasing
assert np.all(
pred_errors.values[1:].round(4) >= pred_errors.values[:-1].round(4)
)
def test_predict_time_index_in_sample_full(Forecaster, index_type, fh_type,
is_relative):
# Check that predicted time index matched forecasting horizon for full in-sample
# predictions.
y_train = make_forecasting_problem(index_type=index_type)
cutoff = y_train.index[-1]
steps = -np.arange(len(y_train)) # full in-sample fh
fh = _make_fh(cutoff, steps, fh_type, is_relative)
f = _construct_instance(Forecaster)
try:
f.fit(y_train, fh=fh)
y_pred = f.predict()
assert_correct_pred_time_index(y_pred.index, y_train.index[-1], fh)
except NotImplementedError:
pass
def test_update_predict_single(Forecaster, fh):
# Check correct time index of update-predict
f = _construct_instance(Forecaster)
f.fit(y_train, fh=fh)
y_pred = f.update_predict_single(y_test)
assert_correct_pred_time_index(y_pred.index, y_test.index[-1], fh)
def test_update_predict_single(Forecaster, fh):
f = _construct_instance(Forecaster)
f.fit(y_train, fh)
y_pred = f.update_predict_single(y_test)
assert_correct_pred_time_index(y_pred, y_test, fh)
def test_predict_time_index(Forecaster, fh, y_train):
f = _construct_instance(Forecaster)
f.fit(y_train, fh)
y_pred = f.predict()
assert_correct_pred_time_index(y_pred, y_train, fh)
