DOC fix links in truncated SVD docs (scikit-learn#17194)

NicolasHug · adrinjalali · commit 201060fe3210 · 2020-05-19T10:00:50.000+02:00
diff --git a/doc/modules/decomposition.rst b/doc/modules/decomposition.rst
@@ -288,7 +288,8 @@ Truncated singular value decomposition and latent semantic analysis
 where :math:`k` is a user-specified parameter.
 
 When truncated SVD is applied to term-document matrices
-(as returned by ``CountVectorizer`` or ``TfidfVectorizer``),
+(as returned by :class:`~sklearn.feature_extraction.text.CountVectorizer` or
+:class:`~sklearn.feature_extraction.text.TfidfVectorizer`),
 this transformation is known as
 `latent semantic analysis <https://nlp.stanford.edu/IR-book/pdf/18lsi.pdf>`_
 (LSA), because it transforms such matrices
@@ -327,8 +328,7 @@ To also transform a test set :math:`X`, we multiply it with :math:`V_k`:
     but the singular values found are the same.
 
 :class:`TruncatedSVD` is very similar to :class:`PCA`, but differs
-in that it works on sample matrices :math:`X` directly
-instead of their covariance matrices.
+in that the matrix :math:`X` does not need to be centered.
 When the columnwise (per-feature) means of :math:`X`
 are subtracted from the feature values,
 truncated SVD on the resulting matrix is equivalent to PCA.
@@ -338,7 +338,7 @@ matrices without the need to densify them,
 as densifying may fill up memory even for medium-sized document collections.
 
 While the :class:`TruncatedSVD` transformer
-works with any (sparse) feature matrix,
+works with any feature matrix,
 using it on tf–idf matrices is recommended over raw frequency counts
 in an LSA/document processing setting.
 In particular, sublinear scaling and inverse document frequency
diff --git a/sklearn/decomposition/_truncated_svd.py b/sklearn/decomposition/_truncated_svd.py
@@ -27,16 +27,16 @@ class TruncatedSVD(TransformerMixin, BaseEstimator):
     This transformer performs linear dimensionality reduction by means of
     truncated singular value decomposition (SVD). Contrary to PCA, this
     estimator does not center the data before computing the singular value
-    decomposition. This means it can work with scipy.sparse matrices
+    decomposition. This means it can work with sparse matrices
     efficiently.
 
     In particular, truncated SVD works on term count/tf-idf matrices as
-    returned by the vectorizers in sklearn.feature_extraction.text. In that
-    context, it is known as latent semantic analysis (LSA).
+    returned by the vectorizers in :mod:`sklearn.feature_extraction.text`. In
+    that context, it is known as latent semantic analysis (LSA).
 
     This estimator supports two algorithms: a fast randomized SVD solver, and
-    a "naive" algorithm that uses ARPACK as an eigensolver on (X * X.T) or
-    (X.T * X), whichever is more efficient.
+    a "naive" algorithm that uses ARPACK as an eigensolver on `X * X.T` or
+    `X.T * X`, whichever is more efficient.
 
     Read more in the :ref:`User Guide <LSA>`.
 
@@ -56,8 +56,8 @@ class TruncatedSVD(TransformerMixin, BaseEstimator):
     n_iter : int, optional (default 5)
         Number of iterations for randomized SVD solver. Not used by ARPACK. The
         default is larger than the default in
-        `~sklearn.utils.extmath.randomized_svd` to handle sparse matrices that
-        may have large slowly decaying spectrum.
+        :func:`~sklearn.utils.extmath.randomized_svd` to handle sparse
+        matrices that may have large slowly decaying spectrum.
 
     random_state : int, RandomState instance, default=None
         Used during randomized svd. Pass an int for reproducible results across
