diff --git a/advanced/optimizing/index.rst b/advanced/optimizing/index.rst
index 13c172e26..05e7d43f3 100644
--- a/advanced/optimizing/index.rst
+++ b/advanced/optimizing/index.rst
@@ -93,7 +93,7 @@ Useful when you have a large program to profile, for example the
     (`ICA <https://en.wikipedia.org/wiki/Independent_component_analysis>`_). PCA
     is a technique for dimensionality reduction, i.e. an algorithm to explain
     the observed variance in your data using less dimensions. ICA is a source
-    seperation technique, for example to unmix multiple signals that have been
+    separation technique, for example to unmix multiple signals that have been
     recorded through multiple sensors. Doing a PCA first and then an ICA can be
     useful if you have more sensors than signals. For more information see:
     `the FastICA example from scikits-learn <http://scikit-learn.org/stable/auto_examples/decomposition/plot_ica_blind_source_separation.html>`_.