categorical update

Author: artidoro

src/Microsoft.ML.Data/Transforms/ConversionsExtensionsCatalog.cs

@@ -24,7 +24,7 @@ public static class ConversionsExtensionsCatalog
         /// </summary>
         /// <param name="catalog">The conversion transform's catalog.</param>
         /// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.
-        /// This column's data type will be a vector of <see cref="System.UInt32"/>, or a scalar <see cref="System.UInt32"/> based on whether the input column data types
+        /// This column's data type will be a vector of keys, or a scalar keys based on whether the input column data types
         /// are vectors or scalars.</param>
         /// <param name="inputColumnName">Name of the column whose data will be hashed.
         /// If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/> will be used as source.

src/Microsoft.ML.Data/Transforms/Hashing.cs

@@ -1113,7 +1113,7 @@ public override void Process()
     /// | -- | -- |
     /// | Does this estimator need to look at the data to train its parameters? | Yes, if the mapping of the hashes to the values is required. |
     /// | Input column data type | Vector or scalars of numeric, boolean, [text](xref:Microsoft.ML.Data.TextDataViewType), [DateTime](xref: System.DateTime) and [key](xref:Microsoft.ML.Data.KeyDataViewType) data types.|
-    /// | Output column data type | Vector or scalar [System.Int32](xref:System.Int32).|
+    /// | Output column data type | Vector or scalar [key](xref:Microsoft.ML.Data.KeyDataViewType)|
     ///
     /// ]]></format>
     /// </remarks>

src/Microsoft.ML.Data/Transforms/ValueToKeyMappingEstimator.cs

@@ -21,8 +21,8 @@ namespace Microsoft.ML.Transforms
     /// |  |  |
     /// | -- | -- |
     /// | Does this estimator need to look at the data to train its parameters? | Yes |
-    /// | Input column data type | Scalar numeric, boolean, [text](xref:Microsoft.ML.Data.TextDataViewType), [System.DateTime](xref:System.DateTime) or [key](xref:Microsoft.ML.Data.KeyDataViewType) data types.|
-    /// | Output column data type | [key](xref:Microsoft.ML.Data.KeyDataViewType)|
+    /// | Input column data type | Scalar or vector of numeric, boolean, [text](xref:Microsoft.ML.Data.TextDataViewType), [System.DateTime](xref:System.DateTime) and [key](xref:Microsoft.ML.Data.KeyDataViewType) data types.|
+    /// | Output column data type | Scalar or vector of [key](xref:Microsoft.ML.Data.KeyDataViewType)|
     ///
     /// The ValueToKeyMappingEstimator builds up term vocabularies(dictionaries) mapping the input values to the keys on the dictionary.
     /// If multiple columns are used, each column builds/uses exactly one vocabulary.

src/Microsoft.ML.Transforms/CategoricalCatalog.cs

@@ -15,17 +15,22 @@ namespace Microsoft.ML
     public static class CategoricalCatalog
     {
         /// <summary>
-        /// Create a <see cref="OneHotEncodingEstimator"/>, which converts the text input column specified by <paramref name="inputColumnName"/> into a column of one-hot encoded vectors named <paramref name="outputColumnName"/>.
+        /// Create a <see cref="OneHotEncodingEstimator"/>, which converts the input column specified by <paramref name="inputColumnName"/>
+        /// into a column of one-hot encoded vectors named <paramref name="outputColumnName"/>.
         /// </summary>
-        /// <param name="catalog">The transform catalog</param>
+        /// <param name="catalog">The transform catalog.</param>
         /// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.
-        /// This column's data type will be a vector of floats for <paramref name="outputKind"/> Bag, Indicator, and Binary. For <paramref name="outputKind"/> Key, the data type will be a key in the case of a singleton input column or a vector of keys in the case of a vector input column.</param>
-        /// <param name="inputColumnName">Name of column to convert to one-hot vectors. If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/> will be used as source.
-        /// This column's data type can be numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/></param>
+        /// This column's data type will be a vector of <see cref="System.Single"/> if <paramref name="outputKind"/> is
+        /// <see cref="OneHotEncodingEstimator.OutputKind.Bag"/>, <see cref="OneHotEncodingEstimator.OutputKind.Indicator"/>, and <see cref="OneHotEncodingEstimator.OutputKind.Binary"/>.
+        /// If <paramref name="outputKind"/> is <see cref="OneHotEncodingEstimator.OutputKind.Key"/>, this column's data type will be a key in the case of a scalar input column
+        /// or a vector of keys in the case of a vector input column.</param>
+        /// <param name="inputColumnName">Name of column to convert to one-hot vectors. If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/>
+        /// will be used as source. This column's data type can be scalar or vector of numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/>,</param>
         /// <param name="outputKind">Output kind: Bag (multi-set vector), Indicator (indicator vector), Key (index), or Binary encoded indicator vector.</param>
         /// <param name="maximumNumberOfKeys">Maximum number of terms to keep per column when auto-training.</param>
-        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/> choosen they will be in the order encountered.
-        /// If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>, items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
+        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/>
+        /// choosen they will be in the order encountered. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>,
+        /// items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
         /// <param name="keyData">Specifies an ordering for the encoding. If specified, this should be a single column data view,
         /// and the key-values will be taken from that column. If unspecified, the ordering will be determined from the input data upon fitting.</param>
         /// <example>
@@ -45,16 +50,19 @@ public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.Cate
                 new[] { new OneHotEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, maximumNumberOfKeys, keyOrdinality) }, keyData);
 
         /// <summary>
-        /// Create a <see cref="OneHotEncodingEstimator"/>, which converts the input text column specified by <see cref="InputOutputColumnPair.InputColumnName"/> into a column of one-hot encoded vectors named <see cref="InputOutputColumnPair.OutputColumnName"/>.
+        /// Create a <see cref="OneHotEncodingEstimator"/>, which converts one or more input text columns specified in <paramref name="columns"/>
+        /// into as many columns of one-hot encoded vectors.
         /// </summary>
-        /// <param name="catalog">The transform catalog</param>
-        /// <param name="columns">The pairs of input and output columns. The data type of the input column can be numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/>.
-        /// The data type of the output column will be a vector of floats for <paramref name="outputKind"/> Bag, Indicator, and Binary.
-        /// For <paramref name="outputKind"/> Key, the data type of the output column will be a key in the case of a singleton input column or a vector of keys in the case of a vector input column.</param>
+        /// <param name="catalog">The transform catalog.</param>
+        /// <param name="columns">The pairs of input and output columns. The output columns' data type will be a vector of <see cref="System.Single"/> if <paramref name="outputKind"/> is
+        /// <see cref="OneHotEncodingEstimator.OutputKind.Bag"/>, <see cref="OneHotEncodingEstimator.OutputKind.Indicator"/>, and <see cref="OneHotEncodingEstimator.OutputKind.Binary"/>.
+        /// If <paramref name="outputKind"/> is <see cref="OneHotEncodingEstimator.OutputKind.Key"/>, the output columns' data type will be a key in the case of scalar input column
+        /// or a vector of keys in the case of a vector input column.</param>
         /// <param name="outputKind">Output kind: Bag (multi-set vector), Ind (indicator vector), Key (index), or Binary encoded indicator vector.</param>
         /// <param name="maximumNumberOfKeys">Maximum number of terms to keep per column when auto-training.</param>
-        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/> choosen they will be in the order encountered.
-        /// If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>, items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
+        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/>
+        /// choosen they will be in the order encountered. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>,
+        /// items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
         /// <param name="keyData">Specifies an ordering for the encoding. If specified, this should be a single column data view,
         /// and the key-values will be taken from that column. If unspecified, the ordering will be determined from the input data upon fitting.</param>
         /// <example>
@@ -100,19 +108,24 @@ internal static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.Ca
             => new OneHotEncodingEstimator(CatalogUtils.GetEnvironment(catalog), columns, keyData);
 
         /// <summary>
-        /// Create a <see cref="OneHotHashEncodingEstimator"/>, which converts a text column specified by <paramref name="inputColumnName"/> into a hash-based one-hot encoded vector column named <paramref name="outputColumnName"/>.
+        /// Create a <see cref="OneHotHashEncodingEstimator"/>, which converts a text column specified by <paramref name="inputColumnName"/>
+        /// into a hash-based one-hot encoded vector column named <paramref name="outputColumnName"/>.
         /// </summary>
         /// <param name="catalog">The transform catalog</param>
         /// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.
-        /// This column's data type will be a vector of floats for <paramref name="outputKind"/> Bag, Indicator, and Binary. For <paramref name="outputKind"/> Key, the data type will be a key in the case of a singleton input column or a vector of keys in the case of a vector input column.</param>
+        /// This column's data type will be a vector of <see cref="System.Single"/> if <paramref name="outputKind"/> is
+        /// <see cref="OneHotEncodingEstimator.OutputKind.Bag"/>, <see cref="OneHotEncodingEstimator.OutputKind.Indicator"/>, and <see cref="OneHotEncodingEstimator.OutputKind.Binary"/>.
+        /// If <paramref name="outputKind"/> is <see cref="OneHotEncodingEstimator.OutputKind.Key"/>, this column's data type will be a key in the case of a scalar input column
+        /// or a vector of keys in the case of a vector input column.
         /// <param name="inputColumnName">Name of column to transform. If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/> will be used as source.
-        /// This column's data type can be numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/>.</param>
+        /// This column's data type can be scalar or vector of numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/>.</param>
         /// <param name="outputKind">The conversion mode.</param>
         /// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</param>
         /// <param name="seed">Hashing seed.</param>
         /// <param name="useOrderedHashing">Whether the position of each term should be included in the hash.</param>
         /// <param name="maximumNumberOfInverts">During hashing we constuct mappings between original values and the produced hash values.
-        /// Text representation of original values are stored in the slot names of the  metadata for the new column.Hashing, as such, can map many initial values to one.
+        /// Text representation of original values are stored in the slot names of the  metadata for the new column.Hashing,
+        /// as such, can map many initial values to one.</param>
         /// <paramref name="maximumNumberOfInverts"/> specifies the upper bound of the number of distinct input values mapping to a hash that should be retained.
         /// <value>0</value> does not retain any input values. <value>-1</value> retains all input values mapping to each hash.</param>
         /// <example>
@@ -133,12 +146,14 @@ public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCata
                 new[] { new OneHotHashEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts) });
 
         /// <summary>
-        /// Create a <see cref="OneHotHashEncodingEstimator"/>, which converts the input text column specified by <see cref="InputOutputColumnPair.InputColumnName"/> into a column of hash-based one-hot encoded vectors named <see cref="InputOutputColumnPair.OutputColumnName"/>
+        /// Create a <see cref="OneHotHashEncodingEstimator"/>, which converts one or more input text columns specified by <paramref name="columns"/>
+        /// into as many columns of hash-based one-hot encoded vectors.
         /// </summary>
         /// <param name="catalog">The transform catalog</param>
-        /// <param name="columns">The pairs of input and output columns. The data type of the input column can be numeric, text, boolean, <see cref="System.DateTime"/> or <see cref="System.DateTimeOffset"/>.
-        /// The data type of the output column will be a vector of floats for <paramref name="outputKind"/> Bag, Indicator, and Binary.
-        /// For <paramref name="outputKind"/> Key, the data type of the output column will be a key in the case of a singleton input column or a vector of keys in the case of a vector input column.</param>
+        /// <param name="columns">The pairs of input and output columns. The output columns' data type will be a vector of <see cref="System.Single"/> if <paramref name="outputKind"/> is
+        /// <see cref="OneHotEncodingEstimator.OutputKind.Bag"/>, <see cref="OneHotEncodingEstimator.OutputKind.Indicator"/>, and <see cref="OneHotEncodingEstimator.OutputKind.Binary"/>.
+        /// If <paramref name="outputKind"/> is <see cref="OneHotEncodingEstimator.OutputKind.Key"/>, the output columns' data type will be a key in the case of scalar input column
+        /// or a vector of keys in the case of a vector input column.</param>
         /// <param name="outputKind">The conversion mode.</param>
         /// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</param>
         /// <param name="seed">Hashing seed.</param>