private void SetScorer()
{
var schema = new RoleMappedSchema(TrainSchema, LabelColumnName, FeatureColumnName);
var args = new BinaryClassifierScorer.Arguments {
Threshold = Threshold, ThresholdColumn = ThresholdColumn
};
Scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
private void SetScorer()
{
var schema = new RoleMappedSchema(TrainSchema, _trainLabelColumn, FeatureColumnName);
var args = new MulticlassClassificationScorer.Arguments()
{
ScoreColumnName = _scoreColumn, PredictedLabelColumnName = _predictedLabelColumn
};
Scorer = new MulticlassClassificationScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
private void SetScorer()
{
var schema = new RoleMappedSchema(TrainSchema, _trainLabelColumn, FeatureColumn);
var args = new MultiClassClassifierScorer.Arguments();
Scorer = new MultiClassClassifierScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
private protected GenericScorer GetGenericScorer()
{
var schema = new RoleMappedSchema(TrainSchema, null, FeatureColumnName);
return(new GenericScorer(Host, new GenericScorer.Arguments(), new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema));
}
public FieldAwareFactorizationMachinePredictionTransformer(IHostEnvironment host, FieldAwareFactorizationMachinePredictor model, Schema trainSchema,
string[] featureColumns, float threshold = 0f, string thresholdColumn = DefaultColumnNames.Score)
: base(Contracts.CheckRef(host, nameof(host)).Register(nameof(FieldAwareFactorizationMachinePredictionTransformer)), model, trainSchema)
{
Host.CheckNonEmpty(thresholdColumn, nameof(thresholdColumn));
_threshold = threshold;
_thresholdColumn = thresholdColumn;
Host.CheckValue(featureColumns, nameof(featureColumns));
int featCount = featureColumns.Length;
Host.Check(featCount >= 0, "Empty features column.");
FeatureColumns = featureColumns;
FeatureColumnTypes = new ColumnType[featCount];
int i = 0;
foreach (var feat in featureColumns)
{
if (!trainSchema.TryGetColumnIndex(feat, out int col))
{
throw Host.ExceptSchemaMismatch(nameof(featureColumns), RoleMappedSchema.ColumnRole.Feature.Value, feat);
}
FeatureColumnTypes[i++] = trainSchema.GetColumnType(col);
}
BindableMapper = ScoreUtils.GetSchemaBindableMapper(Host, model);
var schema = GetSchema();
var args = new BinaryClassifierScorer.Arguments {
Threshold = _threshold, ThresholdColumn = _thresholdColumn
};
Scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, trainSchema), BindableMapper.Bind(Host, schema), schema);
}
public FieldAwareFactorizationMachinePredictionTransformer(IHostEnvironment host, ModelLoadContext ctx)
: base(Contracts.CheckRef(host, nameof(host)).Register(nameof(FieldAwareFactorizationMachinePredictionTransformer)), ctx)
{
// *** Binary format ***
// <base info>
// ids of strings: feature columns.
// float: scorer threshold
// id of string: scorer threshold column
// count of feature columns. FAFM uses more than one.
int featCount = Model.FieldCount;
FeatureColumns = new string[featCount];
FeatureColumnTypes = new ColumnType[featCount];
for (int i = 0; i < featCount; i++)
{
FeatureColumns[i] = ctx.LoadString();
if (!TrainSchema.TryGetColumnIndex(FeatureColumns[i], out int col))
{
throw Host.ExceptSchemaMismatch(nameof(FeatureColumns), RoleMappedSchema.ColumnRole.Feature.Value, FeatureColumns[i]);
}
FeatureColumnTypes[i] = TrainSchema.GetColumnType(col);
}
_threshold = ctx.Reader.ReadSingle();
_thresholdColumn = ctx.LoadString();
BindableMapper = ScoreUtils.GetSchemaBindableMapper(Host, Model);
var schema = GetSchema();
var args = new BinaryClassifierScorer.Arguments {
Threshold = _threshold, ThresholdColumn = _thresholdColumn
};
Scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
internal RankingPredictionTransformer(IHostEnvironment env, ModelLoadContext ctx)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(RankingPredictionTransformer <TModel>)), ctx)
{
var schema = new RoleMappedSchema(TrainSchema, null, FeatureColumn);
_scorer = new GenericScorer(Host, new GenericScorer.Arguments(), new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
/// <summary>
/// The counter constructor of re-creating <see cref="MatrixFactorizationPredictionTransformer"/> from the context where
/// the original transform is saved.
/// </summary>
public MatrixFactorizationPredictionTransformer(IHostEnvironment host, ModelLoadContext ctx)
: base(Contracts.CheckRef(host, nameof(host)).Register(nameof(MatrixFactorizationPredictionTransformer)), ctx)
{
// *** Binary format ***
// <base info>
// string: the column name of matrix's column ids.
// string: the column name of matrix's row ids.
MatrixColumnIndexColumnName = ctx.LoadString();
MatrixRowIndexColumnName = ctx.LoadString();
if (!TrainSchema.TryGetColumnIndex(MatrixColumnIndexColumnName, out int xCol))
{
throw Host.ExceptSchemaMismatch(nameof(MatrixColumnIndexColumnName), RecommenderUtils.MatrixColumnIndexKind.Value, MatrixColumnIndexColumnName);
}
MatrixColumnIndexColumnType = TrainSchema.GetColumnType(xCol);
if (!TrainSchema.TryGetColumnIndex(MatrixRowIndexColumnName, out int yCol))
{
throw Host.ExceptSchemaMismatch(nameof(MatrixRowIndexColumnName), RecommenderUtils.MatrixRowIndexKind.Value, MatrixRowIndexColumnName);
}
MatrixRowIndexColumnType = TrainSchema.GetColumnType(yCol);
BindableMapper = ScoreUtils.GetSchemaBindableMapper(Host, Model);
var schema = GetSchema();
var args = new GenericScorer.Arguments {
Suffix = ""
};
Scorer = new GenericScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
public MulticlassPredictionTransformer(IHostEnvironment env, ModelLoadContext ctx)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(MulticlassPredictionTransformer <TModel>)), ctx)
{
// *** Binary format ***
// <base info>
// id of string: train label column
_trainLabelColumn = ctx.LoadStringOrNull();
var schema = new RoleMappedSchema(TrainSchema, _trainLabelColumn, FeatureColumn);
var args = new MultiClassClassifierScorer.Arguments();
_scorer = new MultiClassClassifierScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
public RankingPredictionTransformer(IHostEnvironment env, TModel model, ISchema inputSchema, string featureColumn)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(RankingPredictionTransformer <TModel>)), model, inputSchema, featureColumn)
{
var schema = new RoleMappedSchema(inputSchema, null, featureColumn);
_scorer = new GenericScorer(Host, new GenericScorer.Arguments(), new EmptyDataView(Host, inputSchema), BindableMapper.Bind(Host, schema), schema);
}
public MulticlassPredictionTransformer(IHostEnvironment env, TModel model, ISchema inputSchema, string featureColumn, string labelColumn)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(MulticlassPredictionTransformer <TModel>)), model, inputSchema, featureColumn)
{
Host.CheckValueOrNull(labelColumn);
_trainLabelColumn = labelColumn;
var schema = new RoleMappedSchema(inputSchema, labelColumn, featureColumn);
var args = new MultiClassClassifierScorer.Arguments();
_scorer = new MultiClassClassifierScorer(Host, args, new EmptyDataView(Host, inputSchema), BindableMapper.Bind(Host, schema), schema);
}
public BinaryPredictionTransformer(IHostEnvironment env, ModelLoadContext ctx)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(BinaryPredictionTransformer <TModel>)), ctx)
{
// *** Binary format ***
// <base info>
// float: scorer threshold
// id of string: scorer threshold column
Threshold = ctx.Reader.ReadSingle();
ThresholdColumn = ctx.LoadString();
var schema = new RoleMappedSchema(TrainSchema, null, FeatureColumn);
var args = new BinaryClassifierScorer.Arguments {
Threshold = Threshold, ThresholdColumn = ThresholdColumn
};
_scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
public BinaryPredictionTransformer(IHostEnvironment env, TModel model, ISchema inputSchema, string featureColumn,
float threshold = 0f, string thresholdColumn = DefaultColumnNames.Score)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(BinaryPredictionTransformer <TModel>)), model, inputSchema, featureColumn)
{
Host.CheckNonEmpty(thresholdColumn, nameof(thresholdColumn));
var schema = new RoleMappedSchema(inputSchema, null, featureColumn);
Threshold = threshold;
ThresholdColumn = thresholdColumn;
var args = new BinaryClassifierScorer.Arguments {
Threshold = Threshold, ThresholdColumn = ThresholdColumn
};
_scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, inputSchema), BindableMapper.Bind(Host, schema), schema);
}
internal ClusteringPredictionTransformer(IHostEnvironment env, TModel model, DataViewSchema inputSchema, string featureColumn,
float threshold = 0f, string thresholdColumn = DefaultColumnNames.Score)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(ClusteringPredictionTransformer <TModel>)), model, inputSchema, featureColumn)
{
Host.CheckNonEmpty(thresholdColumn, nameof(thresholdColumn));
var schema = new RoleMappedSchema(inputSchema, null, featureColumn);
var args = new ClusteringScorer.Arguments();
Scorer = new ClusteringScorer(Host, args, new EmptyDataView(Host, inputSchema), BindableMapper.Bind(Host, schema), schema);
}
internal FieldAwareFactorizationMachinePredictionTransformer(IHostEnvironment host, FieldAwareFactorizationMachineModelParameters model, DataViewSchema trainSchema,
string[] featureColumns, float threshold = 0f, string thresholdColumn = DefaultColumnNames.Score)
: base(Contracts.CheckRef(host, nameof(host)).Register(nameof(FieldAwareFactorizationMachinePredictionTransformer)), model, trainSchema)
{
Host.CheckNonEmpty(thresholdColumn, nameof(thresholdColumn));
Host.CheckNonEmpty(featureColumns, nameof(featureColumns));
_threshold = threshold;
_thresholdColumn = thresholdColumn;
FeatureColumns = featureColumns;
var featureColumnTypes = new DataViewType[featureColumns.Length];
int i = 0;
foreach (var feat in featureColumns)
{
if (!trainSchema.TryGetColumnIndex(feat, out int col))
{
throw Host.ExceptSchemaMismatch(nameof(featureColumns), "feature", feat);
}
featureColumnTypes[i++] = trainSchema[col].Type;
}
FeatureColumnTypes = featureColumnTypes;
BindableMapper = ScoreUtils.GetSchemaBindableMapper(Host, model);
var schema = GetSchema();
var args = new BinaryClassifierScorer.Arguments {
Threshold = _threshold, ThresholdColumn = _thresholdColumn
};
Scorer = new BinaryClassifierScorer(Host, args, new EmptyDataView(Host, trainSchema), BindableMapper.Bind(Host, schema), schema);
}
internal ClusteringPredictionTransformer(IHostEnvironment env, ModelLoadContext ctx)
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(ClusteringPredictionTransformer <TModel>)), ctx)
{
// *** Binary format ***
// <base info>
var schema = new RoleMappedSchema(TrainSchema, null, FeatureColumnName);
var args = new ClusteringScorer.Arguments();
Scorer = new ClusteringScorer(Host, args, new EmptyDataView(Host, TrainSchema), BindableMapper.Bind(Host, schema), schema);
}
/// <summary>
/// Build a transformer based on matrix factorization predictor (model) and the input schema (trainSchema). The created
/// transformer can only transform IDataView objects compatible to the input schema; that is, that IDataView must contain
/// columns specified by <see cref="MatrixColumnIndexColumnName"/>, <see cref="MatrixColumnIndexColumnType"/>, <see cref="MatrixRowIndexColumnName"/>, and <see cref="MatrixRowIndexColumnType"></see>.
/// The output column is "Score" by default but user can append a string to it.
/// </summary>
/// <param name="env">Eviroment object for showing information</param>
/// <param name="model">The model trained by one of the training functions in <see cref="MatrixFactorizationTrainer"/></param>
/// <param name="trainSchema">Targeted schema that containing columns named as xColumnName</param>
/// <param name="matrixColumnIndexColumnName">The name of the column used as role <see cref="RecommenderUtils.MatrixColumnIndexKind"/> in matrix factorization world</param>
/// <param name="matrixRowIndexColumnName">The name of the column used as role <see cref="RecommenderUtils.MatrixRowIndexKind"/> in matrix factorization world</param>
/// <param name="scoreColumnNameSuffix">A string attached to the output column name of this transformer</param>
public MatrixFactorizationPredictionTransformer(IHostEnvironment env, MatrixFactorizationPredictor model, Schema trainSchema,
string matrixColumnIndexColumnName, string matrixRowIndexColumnName, string scoreColumnNameSuffix = "")
: base(Contracts.CheckRef(env, nameof(env)).Register(nameof(MatrixFactorizationPredictionTransformer)), model, trainSchema)
{
Host.CheckNonEmpty(matrixColumnIndexColumnName, nameof(matrixRowIndexColumnName));
Host.CheckNonEmpty(matrixColumnIndexColumnName, nameof(matrixRowIndexColumnName));
MatrixColumnIndexColumnName = matrixColumnIndexColumnName;
MatrixRowIndexColumnName = matrixRowIndexColumnName;
if (!trainSchema.TryGetColumnIndex(MatrixColumnIndexColumnName, out int xCol))
{
throw Host.ExceptSchemaMismatch(nameof(MatrixColumnIndexColumnName), RecommenderUtils.MatrixColumnIndexKind.Value, MatrixColumnIndexColumnName);
}
MatrixColumnIndexColumnType = trainSchema.GetColumnType(xCol);
if (!trainSchema.TryGetColumnIndex(MatrixRowIndexColumnName, out int yCol))
{
throw Host.ExceptSchemaMismatch(nameof(yCol), RecommenderUtils.MatrixRowIndexKind.Value, MatrixRowIndexColumnName);
}
BindableMapper = ScoreUtils.GetSchemaBindableMapper(Host, model);
var schema = GetSchema();
var args = new GenericScorer.Arguments {
Suffix = scoreColumnNameSuffix
};
Scorer = new GenericScorer(Host, args, new EmptyDataView(Host, trainSchema), BindableMapper.Bind(Host, schema), schema);
}