[ENH] Implement TS-QUAD configuration option in ElasticEnsemble

aeon/classification/distance_based/_elastic_ensemble.py

@@ -42,7 +42,9 @@ class ElasticEnsemble(BaseClassifier):
       A ``list`` of strings identifying which distance measures to include. Valid values
       are one or more of: ``euclidean``, ``dtw``, ``wdtw``, ``ddtw``, ``wddtw``,
       ``lcss``, ``erp``, ``msm``, ``twe``. The default value ``all`` means that all
-      the previously listed distances are used.
+      the previously listed distances are used. The special value ``ts-quad`` can be
+      used to select the distance measures for the TS-QUAD ensemble: WDTW, DDTW,
+      LCSS, and MSM.
     proportion_of_param_options : float, default=1
       The proportion of the parameter grid space to search optional.
     proportion_train_in_param_finding : float, default=1
@@ -153,6 +155,17 @@ def _fit(self, X, y):
                 "euclidean",
                 "twe",
             ]
+        elif self.distance_measures == "ts-quad":
+            self._distance_measures = [
+                "wdtw",
+                "ddtw",
+                "lcss",
+                "msm",
+            ]
+            if self.verbose > 0:
+                print(  # noqa: T201
+                    "Configuring ElasticEnsemble as TS-QUAD with WDTW, DDTW, LCSS, MSM."
+                )
         else:
             self._distance_measures = self.distance_measures
 
@@ -515,6 +528,13 @@ def _get_test_params(cls, parameter_set: str = "default") -> dict | list[dict]:
                 "majority_vote": True,
                 "distance_measures": ["dtw", "ddtw", "wdtw"],
             }
+        elif parameter_set == "ts-quad":
+            return {
+                "proportion_of_param_options": 0.01,
+                "proportion_train_for_test": 0.1,
+                "majority_vote": True,
+                "distance_measures": "ts-quad",
+            }
         else:
             return {
                 "proportion_of_param_options": 0.01,

aeon/classification/distance_based/tests/test_elastic_ensemble.py

@@ -90,3 +90,17 @@ def test_all_distance_measures():
     ee.fit(X, y)
     distances = list(ee.get_metric_params())
     assert len(distances) == 9
+
+
+def test_ts_quad_distance_measures():
+    """Test the 'ts-quad' option of the distance_measures parameter."""
+    X = np.random.random(size=(10, 1, 10))
+    y = np.array([0, 0, 0, 0, 0, 1, 1, 1, 1, 1])
+    ee = ElasticEnsemble(
+        distance_measures="ts-quad",
+        proportion_train_in_param_finding=0.2,
+        proportion_of_param_options=0.1,
+    )
+    ee.fit(X, y)
+    actual_distances = list(ee.get_metric_params())
+    assert len(actual_distances) == 4

docs/changelogs/v1.3.md

@@ -35,6 +35,7 @@ September 2025
 
 ### Enhancements
 
+- [ENH] Implement TS-QUAD as a distance parameter for `ElasticEnsemble` ({pr}`3126`) {user}`Nithurshen`
 - [ENH] Improvements to ST transformer and classifier ({pr}`2968`) {user}`MatthewMiddlehurst`
 - [ENH] KNN n_jobs and updated kneighbours method ({pr}`2578`) {user}`chrisholder`
 - [ENH] Refactor signature code ({pr}`2943`) {user}`TonyBagnall`