[ENH] HierarchyEnsembleForecaster for level- or node-wise application of forecasters on panel/hierarchical data

.all-contributorsrc

@@ -156,15 +156,6 @@
         "design"
       ]
     },
-    {
-      "login": "nilesh05apr",
-      "name": "Nilesh Kumar",
-      "avatar_url": "https://avatars.githubusercontent.com/u/65773314?v=4",
-      "profile": "https://github.com/nilesh05apr",
-      "contributions": [
-        "code"
-      ]
-    },
     {
       "login": "MatthewMiddlehurst",
       "name": "Matthew Middlehurst",
@@ -1933,6 +1924,17 @@
         "example"
       ]
     },
+    {
+      "login": "VyomkeshVyas",
+      "name": "Vyomkesh Vyas",
+      "profile": "https://github.com/VyomkeshVyas",
+      "contributions": [
+        "code",
+        "doc",
+        "example",
+        "test"
+      ]
+    },
     {
       "login": "xxl4tomxu98",
       "name": "Tom Xu",
@@ -1962,7 +1964,16 @@
         "bug"
       ]
     },
-   {
+    {
+      "login": "nilesh05apr",
+      "name": "Nilesh Kumar",
+      "avatar_url": "https://avatars.githubusercontent.com/u/65773314?v=4",
+      "profile": "https://github.com/nilesh05apr",
+      "contributions": [
+        "code"
+      ]
+    },
+    {
       "login": "JonathanBechtel",
       "name": "JonathanBechtel",
       "avatar_url": "https://avatars.githubusercontent.com/u/481696?v=4",

docs/source/api_reference/forecasting.rst

@@ -49,6 +49,7 @@ Pipelines can also be constructed using ``*``, ``+``, and ``|`` dunders.
     ForecastX
     ForecastByLevel
     Permute
+    HierarchyEnsembleForecaster
 
 Reduction
 ---------

sktime/forecasting/compose/__init__.py

@@ -6,6 +6,7 @@
 __author__ = ["mloning"]
 
 __all__ = [
+    "HierarchyEnsembleForecaster",
     "ColumnEnsembleForecaster",
     "EnsembleForecaster",
     "AutoEnsembleForecaster",
@@ -35,6 +36,7 @@
     EnsembleForecaster,
 )
 from sktime.forecasting.compose._grouped import ForecastByLevel
+from sktime.forecasting.compose._hierarchy_ensemble import HierarchyEnsembleForecaster
 from sktime.forecasting.compose._multiplexer import MultiplexForecaster
 from sktime.forecasting.compose._pipeline import (
     ForecastingPipeline,