Outcome: a dict works for any keyword, e.g. weights of each response

CHANGELOG.md

@@ -11,7 +11,9 @@
   `Outcome(['y1', 'y2'], offset={'y1': 'O1', 'y2': 'O2'})` gives the
   columns `y1`, `y2`, `offset[y1]` and `offset[y2]`, in the order of the
   responses; a missing or unknown key raises. A single column,
-  `offset='O'`, is still one column, the same for every response.
+  `offset='O'`, is still one column, the same for every response. The
+  dict works for any keyword, e.g. `weight={'y1': 'w1', 'y2': 'w2'}` for
+  weights of each response.
 
 ## 0.3.0 (2026-10-26)
 

docs/responses.md

@@ -206,7 +206,8 @@ A multi-response model, with a linear predictor for each response,
 takes an offset for each: a dict, keyed by the responses (the columns
 of a list of columns, or the classes of a `Multinomial`, whatever its
 output), gives the columns `offset[...]` in the order of the responses.
-A single column is the same offset for every response.
+A single column is the same offset for every response. The dict works
+for any keyword, e.g. `weight={...}` for weights of each response.
 
 ```{code-cell} ipython3
 Carseats['O'] = 0.5

modelmatrix/outcome.py

@@ -95,6 +95,8 @@ class Outcome(Response):
         `offset={'y1': 'O1', 'y2': 'O2'}` gives the columns
         `offset[y1]` and `offset[y2]`, in the order of the responses.
         A single column, `offset='O'`, is the same for every response.
+        Any keyword can be a dict, e.g. `weight=` for weights of each
+        response.
 
     Notes
     -----

tests/test_outcome.py

@@ -210,3 +210,10 @@ def test_offset_for_each_response_polars(M):
     y = Outcome(['y1', 'y2'], offset={'y1': 'O1', 'y2': 'O2'}).fit_transform(pl.from_pandas(M))
     assert isinstance(y, pl.DataFrame)
     assert y.columns == ['y1', 'y2', 'offset[y1]', 'offset[y2]']
+
+
+def test_weights_for_each_response(M):
+    W = M.assign(w1=1.0, w2=0.5)
+    y = Outcome(['y1', 'y2'], offset='O', weight={'y1': 'w1', 'y2': 'w2'}).fit_transform(W)
+    assert list(y.columns) == ['y1', 'y2', 'offset', 'weight[y1]', 'weight[y2]']
+    np.testing.assert_allclose(y['weight[y2]'], 0.5)