[MRG+1] GaussianProcessRegressor: faster prediction of std (scikit-learn#8591)

hbertrand · Joan Massich · commit 9ef6a6641817 · 2017-04-26T10:57:18.000+02:00
diff --git a/doc/whats_new.rst b/doc/whats_new.rst
@@ -154,6 +154,9 @@ Enhancements
    - Add ``sample_weight`` parameter to :func:`metrics.cohen_kappa_score` by
      Victor Poughon.
 
+   - In :class:`gaussian_process.GaussianProcessRegressor`, method ``predict`` 
+     is a lot faster with ``return_std=True`` by :user:`Hadrien Bertrand <hbertrand>`.
+
 Bug fixes
 .........
    - Fixed a bug where :class:`sklearn.ensemble.IsolationForest` uses an
diff --git a/sklearn/gaussian_process/gpr.py b/sklearn/gaussian_process/gpr.py
@@ -312,7 +312,7 @@ def predict(self, X, return_std=False, return_cov=False):
                 K_inv = L_inv.dot(L_inv.T)
                 # Compute variance of predictive distribution
                 y_var = self.kernel_.diag(X)
-                y_var -= np.einsum("ki,kj,ij->k", K_trans, K_trans, K_inv)
+                y_var -= np.einsum("ij,ij->i", np.dot(K_trans, K_inv), K_trans)
 
                 # Check if any of the variances is negative because of
                 # numerical issues. If yes: set the variance to 0.
