/// <summary>
/// Convert a text column into hash-based one-hot encoded vector.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <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.</param>
/// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</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.
/// <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>
/// <param name="outputKind">The conversion mode.</param>
public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCatalog.CategoricalTransforms catalog,
string outputColumnName,
string inputColumnName = null,
int numberOfBits = OneHotHashEncodingEstimator.Defaults.NumberOfBits,
int maximumNumberOfInverts = OneHotHashEncodingEstimator.Defaults.MaximumNumberOfInverts,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.OutputKind.Indicator)
=> new OneHotHashEncodingEstimator(CatalogUtils.GetEnvironment(catalog), outputColumnName, inputColumnName ?? outputColumnName, numberOfBits, maximumNumberOfInverts, outputKind);
/// <summary>
/// Convert text columns into one-hot encoded vectors.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <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.</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="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>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.CategoricalTransforms catalog,
string outputColumnName,
string inputColumnName = null,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.Defaults.OutKind,
int maximumNumberOfKeys = ValueToKeyMappingEstimator.Defaults.MaximumNumberOfKeys,
ValueToKeyMappingEstimator.KeyOrdinality keyOrdinality = ValueToKeyMappingEstimator.Defaults.Ordinality,
IDataView keyData = null)
=> new OneHotEncodingEstimator(CatalogUtils.GetEnvironment(catalog),
new[] { new OneHotEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, maximumNumberOfKeys, keyOrdinality) }, keyData);
/// <summary>
/// Instantiates a new instance of <see cref="OneHotHashEncodingEstimator"/>.
/// </summary>
/// <param name="env">Host Environment.</param>
/// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <param name="inputColumnName">Name of the column to transform.
/// If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/> will be used as source.</param>
/// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</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.
/// <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>
/// <param name="outputKind">The type of output expected.</param>
internal OneHotHashEncodingEstimator(IHostEnvironment env,
string outputColumnName,
string inputColumnName = null,
int numberOfBits = Defaults.NumberOfBits,
int maximumNumberOfInverts = Defaults.MaximumNumberOfInverts,
OneHotEncodingEstimator.OutputKind outputKind = Defaults.OutputKind)
: this(env, new ColumnOptions(outputColumnName, inputColumnName ?? outputColumnName, outputKind, numberOfBits, maximumNumberOfInverts : maximumNumberOfInverts))
{
}
/// <summary>
/// Convert a text column into hash-based one-hot encoded vector.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <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.</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.
/// <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>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCatalog.CategoricalTransforms catalog,
string outputColumnName,
string inputColumnName = null,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.OutputKind.Indicator,
int numberOfBits = OneHotHashEncodingEstimator.Defaults.NumberOfBits,
uint seed = OneHotHashEncodingEstimator.Defaults.Seed,
bool useOrderedHashing = OneHotHashEncodingEstimator.Defaults.UseOrderedHashing,
int maximumNumberOfInverts = OneHotHashEncodingEstimator.Defaults.MaximumNumberOfInverts)
=> new OneHotHashEncodingEstimator(CatalogUtils.GetEnvironment(catalog),
new[] { new OneHotHashEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts) });
/// <summary>
/// Describes how the transformer handles one column pair.
/// </summary>
/// <param name="name">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <param name="inputColumnName">Name of column to transform. If set to <see langword="null"/>, the value of the <paramref name="name"/> will be used as source.</param>
/// <param name="outputKind">Kind of output: bag, indicator vector etc.</param>
/// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 31, 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.
/// <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>
public ColumnOptions(string name, string inputColumnName = null,
OneHotEncodingEstimator.OutputKind outputKind = Defaults.OutputKind,
int numberOfBits = Defaults.NumberOfBits,
uint seed = Defaults.Seed,
bool useOrderedHashing = Defaults.UseOrderedHashing,
int maximumNumberOfInverts = Defaults.MaximumNumberOfInverts)
{
HashingOptions = new HashingEstimator.ColumnOptions(name, inputColumnName ?? name, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts);
OutputKind = outputKind;
}
/// <summary>
/// A helper method to create <see cref="OneHotHashEncodingTransformer"/>.
/// </summary>
/// <param name="env">Host Environment.</param>
/// <param name="input">Input <see cref="IDataView"/>. This is the output from previous transform or loader.</param>
/// <param name="name">Name of the output column.</param>
/// <param name="source">Name of the column to be transformed. If this is null '<paramref name="name"/>' will be used.</param>
/// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</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.
/// <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>
/// <param name="outputKind">The type of output expected.</param>
private static IDataView Create(IHostEnvironment env,
IDataView input,
string name,
string source = null,
int numberOfBits = OneHotHashEncodingEstimator.Defaults.NumberOfBits,
int maximumNumberOfInverts = OneHotHashEncodingEstimator.Defaults.MaximumNumberOfInverts,
OneHotEncodingEstimator.OutputKind outputKind = OneHotHashEncodingEstimator.Defaults.OutputKind)
{
return(new OneHotHashEncodingEstimator(env, name, source, numberOfBits, maximumNumberOfInverts, outputKind).Fit(input).Transform(input) as IDataView);
}
/// <summary>
/// Convert text columns into one-hot encoded vectors.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="columns">Specifies the names of the columns on which to apply the transformation.</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="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>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.CategoricalTransforms catalog,
InputOutputColumnPair[] columns,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.Defaults.OutKind,
int maximumNumberOfKeys = ValueToKeyMappingEstimator.Defaults.MaximumNumberOfKeys,
ValueToKeyMappingEstimator.KeyOrdinality keyOrdinality = ValueToKeyMappingEstimator.Defaults.Ordinality,
IDataView keyData = null)
{
var env = CatalogUtils.GetEnvironment(catalog);
env.CheckValue(columns, nameof(columns));
var columnOptions = columns.Select(x => new OneHotEncodingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, outputKind, maximumNumberOfKeys, keyOrdinality)).ToArray();
return(new OneHotEncodingEstimator(env, columnOptions, keyData));
}
public static IEstimator <ITransformer> _OneHotHashEncoding(this MLContext MLContext, JToken componentObject)
{
string outputColumn = componentObject.Value <string>("OutputColumnName");
string inputColumn = componentObject.Value <string>("InputColumnName");
OneHotEncodingEstimator.OutputKind outputKind = Enum.Parse <OneHotEncodingEstimator.OutputKind>(componentObject.Value <string>("OutputKind"));
int numBits = componentObject.Value <int>("NumberOfBits");
uint seed = componentObject.Value <uint>("Seed");
bool useOrderedHashing = componentObject.Value <bool>("UseOrderedHashing");
int maxNumInverts = componentObject.Value <int>("MaximumNumberOfInverts");
return(MLContext.Transforms.Categorical.OneHotHashEncoding(outputColumn, inputColumn, outputKind, numBits,
seed, useOrderedHashing, maxNumInverts));
}
/// <summary>
/// Convert text columns into hash-based one-hot encoded vector columns.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="columns">Specifies the names of the columns on which to apply the transformation.</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.
/// <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>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCatalog.CategoricalTransforms catalog,
InputOutputColumnPair[] columns,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.OutputKind.Indicator,
int numberOfBits = OneHotHashEncodingEstimator.Defaults.NumberOfBits,
uint seed = OneHotHashEncodingEstimator.Defaults.Seed,
bool useOrderedHashing = OneHotHashEncodingEstimator.Defaults.UseOrderedHashing,
int maximumNumberOfInverts = OneHotHashEncodingEstimator.Defaults.MaximumNumberOfInverts)
{
var env = CatalogUtils.GetEnvironment(catalog);
env.CheckValue(columns, nameof(columns));
var columnOptions = columns.Select(x => new OneHotHashEncodingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, outputKind, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts)).ToArray();
return(new OneHotHashEncodingEstimator(env, columnOptions));
}
internal OneHotHashEncodingEstimator(IHostEnvironment env, params ColumnOptions[] columns)
{
Contracts.CheckValue(env, nameof(env));
_host = env.Register(nameof(ValueToKeyMappingEstimator));
_hash = new HashingEstimator(_host, columns.Select(x => x.HashingOptions).ToArray());
using (var ch = _host.Start(nameof(OneHotHashEncodingEstimator)))
{
var binaryCols = new List <(string outputColumnName, string inputColumnName)>();
var cols = new List <(string outputColumnName, string inputColumnName, bool bag)>();
for (int i = 0; i < columns.Length; i++)
{
var column = columns[i];
OneHotEncodingEstimator.OutputKind kind = columns[i].OutputKind;
switch (kind)
{
default:
throw _host.ExceptUserArg(nameof(column.OutputKind));
case OneHotEncodingEstimator.OutputKind.Key:
continue;
case OneHotEncodingEstimator.OutputKind.Binary:
if ((column.HashingOptions.MaximumNumberOfInverts) != 0)
{
ch.Warning("Invert hashing is being used with binary encoding.");
}
binaryCols.Add((column.HashingOptions.Name, column.HashingOptions.Name));
break;
case OneHotEncodingEstimator.OutputKind.Indicator:
cols.Add((column.HashingOptions.Name, column.HashingOptions.Name, false));
break;
case OneHotEncodingEstimator.OutputKind.Bag:
cols.Add((column.HashingOptions.Name, column.HashingOptions.Name, true));
break;
}
}
IEstimator <ITransformer> toBinVector = null;
IEstimator <ITransformer> toVector = null;
if (binaryCols.Count > 0)
{
toBinVector = new KeyToBinaryVectorMappingEstimator(_host, binaryCols.Select(x => (x.outputColumnName, x.inputColumnName)).ToArray());
}
if (cols.Count > 0)
{
toVector = new KeyToVectorMappingEstimator(_host, cols.Select(x => new KeyToVectorMappingEstimator.ColumnOptions(x.outputColumnName, x.inputColumnName, x.bag)).ToArray());
}
if (toBinVector != null && toVector != null)
{
_toSomething = toVector.Append(toBinVector);
}
else
{
if (toBinVector != null)
{
_toSomething = toBinVector;
}
else
{
_toSomething = toVector;
}
}
}
}
/// <summary>
/// Convert a text column into one-hot encoded vector.
/// </summary>
/// <param name="catalog">The transform catalog</param>
/// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
/// <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.</param>
/// <param name="outputKind">The conversion mode.</param>
/// <example>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.CategoricalTransforms catalog,
string outputColumnName,
string inputColumnName = null,
OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.OutputKind.Indicator)
=> new OneHotEncodingEstimator(CatalogUtils.GetEnvironment(catalog), outputColumnName, inputColumnName, outputKind);
