scikit-learn · OmarManzoor · Aug 10, 2023 · Mar 29, 2023 · Mar 29, 2023 · Mar 31, 2023
diff --git a/doc/whats_new/v1.3.rst b/doc/whats_new/v1.3.rst
@@ -619,6 +619,13 @@ Changelog
   `n_targets`, which is used to decide the number of outputs when sampling
   from the prior distributions. :pr:`23099` by :user:`Zhehao Liu <MaxwellLZH>`.
 
+:mod:`sklearn.mixture`
+......................
+
+- |Efficiency| :class:`GaussianMixture` is more efficient now and will bypass unnecessary
+  initialization if the weights, means, and precisions are given by users.
+  :pr:`26021` by :user:`Jiawei Zhang <jiawei-zhang-a>`.
+
 :mod:`sklearn.model_selection`
 ..............................
 

diff --git a/sklearn/mixture/_gaussian_mixture.py b/sklearn/mixture/_gaussian_mixture.py
@@ -754,6 +754,19 @@ def _check_parameters(self, X):
                 n_features,
             )
 
+    def _initialize_parameters(self, X, random_state):
+        # If all the initial parameters are all provided, then there is no need to run
+        # the initialization.
+        compute_resp = (
+            self.weights_init is None
+            or self.means_init is None
+            or self.precisions_init is None
+        )
+        if compute_resp:
+            super()._initialize_parameters(X, random_state)
+        else:
+            self._initialize(X, None)
+
     def _initialize(self, X, resp):
         """Initialization of the Gaussian mixture parameters.
 
@@ -764,11 +777,13 @@ def _initialize(self, X, resp):
         resp : array-like of shape (n_samples, n_components)
         """
         n_samples, _ = X.shape
-
-        weights, means, covariances = _estimate_gaussian_parameters(
-            X, resp, self.reg_covar, self.covariance_type
-        )
-        weights /= n_samples
+        weights, means, covariances = None, None, None
+        if resp is not None:
+            weights, means, covariances = _estimate_gaussian_parameters(
+                X, resp, self.reg_covar, self.covariance_type
+            )
+            if self.weights_init is None:
+                weights /= n_samples
 
         self.weights_ = weights if self.weights_init is None else self.weights_init
         self.means_ = means if self.means_init is None else self.means_init

diff --git a/sklearn/mixture/tests/test_gaussian_mixture.py b/sklearn/mixture/tests/test_gaussian_mixture.py
@@ -8,11 +8,13 @@
 import sys
 import warnings
 from io import StringIO
+from unittest.mock import Mock
 
 import numpy as np
 import pytest
 from scipy import linalg, stats
 
+import sklearn
 from sklearn.cluster import KMeans
 from sklearn.covariance import EmpiricalCovariance
 from sklearn.datasets import make_spd_matrix
@@ -1387,3 +1389,34 @@ def test_gaussian_mixture_single_component_stable():
     X = rng.multivariate_normal(np.zeros(2), np.identity(2), size=3)
     gm = GaussianMixture(n_components=1)
     gm.fit(X).sample()
+
+
+def test_gaussian_mixture_all_init_does_not_estimate_gaussian_parameters(
+    monkeypatch,
+    global_random_seed,
+):
+    """When all init parameters are provided, the Gaussian parameters
+    are not estimated.
+
+    Non-regression test for gh-26015.
+    """
+
+    mock = Mock(side_effect=_estimate_gaussian_parameters)
+    monkeypatch.setattr(
+        sklearn.mixture._gaussian_mixture, "_estimate_gaussian_parameters", mock
+    )
+
+    rng = np.random.RandomState(global_random_seed)
+    rand_data = RandomData(rng)
+
+    gm = GaussianMixture(
+        n_components=rand_data.n_components,
+        weights_init=rand_data.weights,
+        means_init=rand_data.means,
+        precisions_init=rand_data.precisions["full"],
+        random_state=rng,
+    )
+    gm.fit(rand_data.X["full"])
+    # The initial gaussian parameters are not estimated. They are estimated for every
+    # m_step.
+    assert mock.call_count == gm.n_iter_