[SPARK-23127][DOC] Update FeatureHasher guide for categoricalCols parameter

Update user guide entry for `FeatureHasher` to match the Scala / Python doc, to describe the `categoricalCols` parameter.

## How was this patch tested?

Doc only

Author: Nick Pentreath <[email protected]>

Closes #20293 from MLnick/SPARK-23127-catCol-userguide.

(cherry picked from commit 60203fc)
Signed-off-by: Nick Pentreath <[email protected]>

Author: Nick Pentreath

docs/ml-features.md

@@ -222,9 +222,9 @@ The `FeatureHasher` transformer operates on multiple columns. Each column may co
 numeric or categorical features. Behavior and handling of column data types is as follows:
 
 - Numeric columns: For numeric features, the hash value of the column name is used to map the
-feature value to its index in the feature vector. Numeric features are never treated as
-categorical, even when they are integers. You must explicitly convert numeric columns containing
-categorical features to strings first.
+feature value to its index in the feature vector. By default, numeric features are not treated
+as categorical (even when they are integers). To treat them as categorical, specify the relevant
+columns using the `categoricalCols` parameter.
 - String columns: For categorical features, the hash value of the string "column_name=value"
 is used to map to the vector index, with an indicator value of `1.0`. Thus, categorical features
 are "one-hot" encoded (similarly to using [OneHotEncoder](ml-features.html#onehotencoder) with