[ENH] enhance naive forecaster with new strategies (#2869)

* naive forecaster with new strategies

* few modifications according to reviews

* made modifications and add documentation

* modifications according to reviews

* try to pass the test

* delete one comment

* modify 1)more test cases for seasonal_last 2)raise exception for y input in _predict, and related documentation

* fix typo

* take y

* Update _naive.py

---------

Co-authored-by: Tony Bagnall <a.j.bagnall@soton.ac.uk>
Co-authored-by: MatthewMiddlehurst <pfm15hbu@gmail.com>

Author: 3 people

aeon/forecasting/_naive.py

@@ -1,27 +1,79 @@
-"""Naive Forecaster."""
+"""Naive forecaster with multiple strategies."""
+
+__maintainer__ = []
+__all__ = ["NaiveForecaster"]
+
+
+import numpy as np
 
 from aeon.forecasting.base import BaseForecaster
 
 
 class NaiveForecaster(BaseForecaster):
-    """Naive forecaster that always predicts the last value seen in training."""
+    """
+    Naive forecaster with multiple strategies and flexible horizon.
+
+    Parameters
+    ----------
+    strategy : str, default="last"
+        The forecasting strategy to use.
+        Options: "last", "mean", "seasonal_last".
+            - "last" predicts the last value of the input series for all horizon steps.
+            - "mean": predicts the mean of the input series for all horizon steps.
+            - "seasonal_last": predicts the last season value in the training series.
+              Returns np.nan if the effective seasonal data is empty.
+    seasonal_period : int, default=1
+        The seasonal period to use for the "seasonal_last" strategy.
+        E.g., 12 for monthly data with annual seasonality.
+    horizon : int, default =1
+        The number of time steps ahead to forecast. If horizon is one, the forecaster
+        will learn to predict one point ahead.
+        Only relevant for "seasonal_last".
+    """
 
-    def __init__(self):
-        """Initialize NaiveForecaster."""
-        self.last_value_ = None
-        super().__init__(horizon=1, axis=1)
+    def __init__(self, strategy="last", seasonal_period=1, horizon=1):
+        self.strategy = strategy
+        self.seasonal_period = seasonal_period
+
+        super().__init__(horizon=horizon, axis=1)
 
     def _fit(self, y, exog=None):
-        """Fit Naive forecaster."""
-        y = y.squeeze()
-        self.last_value_ = y[-1]
+        y_squeezed = y.squeeze()
+
+        if self.strategy == "last":
+            self._fitted_scalar_value = y_squeezed[-1]
+        elif self.strategy == "mean":
+            self._fitted_scalar_value = np.mean(y_squeezed)
+        elif self.strategy == "seasonal_last":
+            self._fitted_last_season = y_squeezed[-self.seasonal_period :]
+        else:
+            raise ValueError(
+                f"Unknown strategy: {self.strategy}. "
+                "Valid strategies are 'last', 'mean', 'seasonal_last'."
+            )
         return self
 
     def _predict(self, y=None, exog=None):
-        """Predict using Naive forecaster."""
-        return self.last_value_
+        if y is None:
+            if self.strategy == "last" or self.strategy == "mean":
+                return self._fitted_scalar_value
+
+            # For "seasonal_last" strategy
+            prediction_index = (self.horizon - 1) % self.seasonal_period
+            return self._fitted_last_season[prediction_index]
+        else:
+            y_squeezed = y.squeeze()
 
-    def _forecast(self, y, exog=None):
-        """Forecast using dummy forecaster."""
-        y = y.squeeze()
-        return y[-1]
+            if self.strategy == "last":
+                return y_squeezed[-1]
+            elif self.strategy == "mean":
+                return np.mean(y_squeezed)
+            elif self.strategy == "seasonal_last":
+                period = y_squeezed[-self.seasonal_period :]
+                idx = (self.horizon - 1) % self.seasonal_period
+                return period[idx]
+            else:
+                raise ValueError(
+                    f"Unknown strategy: {self.strategy}. "
+                    "Valid strategies are 'last', 'mean', 'seasonal_last'."
+                )

aeon/forecasting/tests/test_naive.py

@@ -0,0 +1,66 @@
+"""Test Naive Forecaster."""
+
+import numpy as np
+
+from aeon.forecasting import NaiveForecaster
+
+
+def test_naive_forecaster_last_strategy():
+    """Test NaiveForecaster with 'last' strategy."""
+    sample_data = np.array([10, 20, 30, 40, 50])
+    forecaster = NaiveForecaster(strategy="last", horizon=3)
+    forecaster.fit(sample_data)
+    predictions = forecaster.predict()
+    expected = 50
+    np.testing.assert_array_equal(predictions, expected)
+
+
+def test_naive_forecaster_mean_strategy():
+    """Test NaiveForecaster with 'mean' strategy."""
+    sample_data = np.array([10, 20, 30, 40, 50])
+    forecaster = NaiveForecaster(strategy="mean", horizon=2)
+    forecaster.fit(sample_data)
+    predictions = forecaster.predict()
+    expected = 30  # Mean of [10, 20, 30, 40, 50] is 30
+    np.testing.assert_array_equal(predictions, expected)
+
+
+def test_naive_forecaster_seasonal_last_strategy():
+    """Test NaiveForecaster with 'seasonal_last' strategy."""
+    data = np.array([1, 2, 3, 4, 5, 6, 7, 8])
+
+    # Last season is [6, 7, 8] for seasonal_period = 3
+    forecaster = NaiveForecaster(strategy="seasonal_last", seasonal_period=3, horizon=4)
+    forecaster.fit(data)
+    pred = forecaster.predict()
+    pred2 = forecaster.predict(y=data)
+    expected = 6  # predicts the 1-st element of the last season.
+    np.testing.assert_array_equal(pred, expected)
+    np.testing.assert_array_equal(pred2, expected)
+
+    # Test horizon within the season length
+    forecaster = NaiveForecaster(strategy="seasonal_last", seasonal_period=3, horizon=2)
+    forecaster.fit(data)
+    pred = forecaster.predict()
+    pred2 = forecaster.predict(y=data)
+    expected = 7  # predicts the 2-nd element of the last season.
+    np.testing.assert_array_equal(pred, expected)
+    np.testing.assert_array_equal(pred2, expected)
+
+    # Test horizon wrapping around to a new season
+    forecaster = NaiveForecaster(strategy="seasonal_last", seasonal_period=3, horizon=7)
+    forecaster.fit(data)
+    pred = forecaster.predict()
+    pred2 = forecaster.predict(y=data)
+    expected = 6  # predicts the 1-st element of the last season.
+    np.testing.assert_array_equal(pred, expected)
+    np.testing.assert_array_equal(pred2, expected)
+
+    # Last season is now [5, 6, 7, 8] with seasonal_period = 4
+    forecaster = NaiveForecaster(strategy="seasonal_last", seasonal_period=4, horizon=6)
+    forecaster.fit(data)
+    pred = forecaster.predict()
+    pred2 = forecaster.predict(y=data)
+    expected = 6  # predicts the 2nd element of the new last season.
+    np.testing.assert_array_equal(pred, expected)
+    np.testing.assert_array_equal(pred2, expected)

docs/api_reference/forecasting.md

@@ -7,8 +7,8 @@
     :toctree: auto_generated/
     :template: class.rst
 
-    DummyForecaster
     BaseForecaster
+    NaiveForecaster
     RegressionForecaster
     ETSForecaster
 ```