/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and log-loss.
/// </summary>
/// <param name="catalog">The binary classification catalog trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="options">Advanced arguments to the algorithm.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), the calibrated prediction (from 0 to 1), and the predicted label.</returns>
/// <example>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static (Scalar <float> score, Scalar <float> probability, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationCatalog.BinaryClassificationTrainers catalog,
Scalar <bool> label, Vector <float> features, Scalar <float> weights,
SdcaBinaryTrainer.Options options,
Action <CalibratedModelParametersBase <LinearBinaryModelParameters, PlattCalibrator> > onFit = null)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValueOrNull(weights);
Contracts.CheckValueOrNull(options);
Contracts.CheckValueOrNull(onFit);
var rec = new TrainerEstimatorReconciler.BinaryClassifier(
(env, labelName, featuresName, weightsName) =>
{
options.LabelColumnName = labelName;
options.FeatureColumnName = featuresName;
var trainer = new SdcaBinaryTrainer(env, options);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
onFit(trans.Model);
}));
}
return(trainer);
}, label, features, weights);
return(rec.Output);
}
/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and log-loss.
/// </summary>
/// <param name="ctx">The binary classification context trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="options">Advanced arguments to the algorithm.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), the calibrated prediction (from 0 to 1), and the predicted label.</returns>
/// <example>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static (Scalar <float> score, Scalar <float> probability, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationContext.BinaryClassificationTrainers ctx,
Scalar <bool> label, Vector <float> features, Scalar <float> weights,
SdcaBinaryTrainer.Options options,
Action <LinearBinaryModelParameters, ParameterMixingCalibratedPredictor> onFit = null)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValueOrNull(weights);
Contracts.CheckValueOrNull(options);
Contracts.CheckValueOrNull(onFit);
var rec = new TrainerEstimatorReconciler.BinaryClassifier(
(env, labelName, featuresName, weightsName) =>
{
options.LabelColumn = labelName;
options.FeatureColumn = featuresName;
var trainer = new SdcaBinaryTrainer(env, options);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
// Under the default log-loss we assume a calibrated predictor.
var model = trans.Model;
var cali = (ParameterMixingCalibratedPredictor)model;
var pred = (LinearBinaryModelParameters)cali.SubPredictor;
onFit(pred, cali);
}));
}
return(trainer);
}, label, features, weights);
return(rec.Output);
}
public void TestEstimatorSymSgdInitPredictor()
{
(var pipe, var dataView) = GetBinaryClassificationPipeline();
var transformedData = pipe.Fit(dataView).Transform(dataView);
var initPredictor = new SdcaBinaryTrainer(Env, "Label", "Features").Fit(transformedData);
var data = initPredictor.Transform(transformedData);
var withInitPredictor = new SymSgdClassificationTrainer(Env, "Label", "Features").Train(transformedData, initialPredictor: initPredictor.Model);
var outInitData = withInitPredictor.Transform(transformedData);
var notInitPredictor = new SymSgdClassificationTrainer(Env, "Label", "Features").Train(transformedData);
var outNoInitData = notInitPredictor.Transform(transformedData);
int numExamples = 10;
var col1 = data.GetColumn <float>(Env, "Score").Take(numExamples).ToArray();
var col2 = outInitData.GetColumn <float>(Env, "Score").Take(numExamples).ToArray();
var col3 = outNoInitData.GetColumn <float>(Env, "Score").Take(numExamples).ToArray();
bool col12Diff = default;
bool col23Diff = default;
bool col13Diff = default;
for (int i = 0; i < numExamples; i++)
{
col12Diff = col12Diff || (col1[i] != col2[i]);
col23Diff = col23Diff || (col2[i] != col3[i]);
col13Diff = col13Diff || (col1[i] != col3[i]);
}
Contracts.Assert(col12Diff && col23Diff && col13Diff);
Done();
}
/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and log-loss.
/// </summary>
/// <param name="catalog">The binary classification catalog trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="l2Const">The L2 regularization hyperparameter.</param>
/// <param name="l1Threshold">The L1 regularization hyperparameter. Higher values will tend to lead to more sparse model.</param>
/// <param name="maxIterations">The maximum number of passes to perform over the data.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), the calibrated prediction (from 0 to 1), and the predicted label.</returns>
/// <example>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static (Scalar <float> score, Scalar <float> probability, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationCatalog.BinaryClassificationTrainers catalog,
Scalar <bool> label, Vector <float> features, Scalar <float> weights = null,
float?l2Const = null,
float?l1Threshold = null,
int?maxIterations = null,
Action <CalibratedModelParametersBase <LinearBinaryModelParameters, PlattCalibrator> > onFit = null)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValueOrNull(weights);
Contracts.CheckParam(!(l2Const < 0), nameof(l2Const), "Must not be negative, if specified.");
Contracts.CheckParam(!(l1Threshold < 0), nameof(l1Threshold), "Must not be negative, if specified.");
Contracts.CheckParam(!(maxIterations < 1), nameof(maxIterations), "Must be positive if specified");
Contracts.CheckValueOrNull(onFit);
var rec = new TrainerEstimatorReconciler.BinaryClassifier(
(env, labelName, featuresName, weightsName) =>
{
var trainer = new SdcaBinaryTrainer(env, labelName, featuresName, weightsName, l2Const, l1Threshold, maxIterations);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
onFit(trans.Model);
}));
}
return(trainer);
}, label, features, weights);
return(rec.Output);
}
public void SetupBreastCancerPipeline()
{
_breastCancerExample = new BreastCancerData()
{
Features = new[] { 5f, 1f, 1f, 1f, 2f, 1f, 3f, 1f, 1f }
};
string _breastCancerDataPath = Program.GetInvariantCultureDataPath("breast-cancer.txt");
var env = new MLContext(seed: 1, conc: 1);
var reader = new TextLoader(env, columns: new[]
{
new TextLoader.Column("Label", DataKind.BL, 0),
new TextLoader.Column("Features", DataKind.R4, new[] { new TextLoader.Range(1, 9) })
},
hasHeader: false
);
IDataView data = reader.Read(_breastCancerDataPath);
var pipeline = new SdcaBinaryTrainer(env, "Label", "Features", advancedSettings: (s) => { s.NumThreads = 1; s.ConvergenceTolerance = 1e-2f; });
var model = pipeline.Fit(data);
_breastCancerModel = model.CreatePredictionEngine <BreastCancerData, BreastCancerPrediction>(env);
}
public void SetupBreastCancerPipeline()
{
_breastCancerExample = new BreastCancerData()
{
Features = new[] { 5f, 1f, 1f, 1f, 2f, 1f, 3f, 1f, 1f }
};
string _breastCancerDataPath = Program.GetInvariantCultureDataPath("breast-cancer.txt");
using (var env = new ConsoleEnvironment(seed: 1, conc: 1, verbose: false, sensitivity: MessageSensitivity.None, outWriter: EmptyWriter.Instance))
{
var reader = new TextLoader(env,
new TextLoader.Arguments()
{
Separator = "\t",
HasHeader = false,
Column = new[]
{
new TextLoader.Column("Label", DataKind.BL, 0),
new TextLoader.Column("Features", DataKind.R4, new[] { new TextLoader.Range(1, 9) })
}
});
IDataView data = reader.Read(_breastCancerDataPath);
var pipeline = new SdcaBinaryTrainer(env, "Label", "Features", advancedSettings: (s) => { s.NumThreads = 1; s.ConvergenceTolerance = 1e-2f; });
var model = pipeline.Fit(data);
_breastCancerModel = model.MakePredictionFunction <BreastCancerData, BreastCancerPrediction>(env);
}
}
private IDataScorerTransform _TrainSentiment()
{
bool normalize = true;
var args = new TextLoader.Arguments()
{
Separator = "tab",
HasHeader = true,
Column = new[] {
new TextLoader.Column("Label", DataKind.BL, 0),
new TextLoader.Column("SentimentText", DataKind.Text, 1)
}
};
var args2 = new TextFeaturizingEstimator.Arguments()
{
Column = new TextFeaturizingEstimator.Column
{
Name = "Features",
Source = new[] { "SentimentText" }
},
KeepDiacritics = false,
KeepPunctuations = false,
TextCase = TextNormalizingEstimator.CaseNormalizationMode.Lower,
OutputTokens = true,
UsePredefinedStopWordRemover = true,
VectorNormalizer = normalize ? TextFeaturizingEstimator.TextNormKind.L2 : TextFeaturizingEstimator.TextNormKind.None,
CharFeatureExtractor = new NgramExtractorTransform.NgramExtractorArguments()
{
NgramLength = 3, AllLengths = false
},
WordFeatureExtractor = new NgramExtractorTransform.NgramExtractorArguments()
{
NgramLength = 2, AllLengths = true
},
};
var trainFilename = FileHelper.GetTestFile("wikipedia-detox-250-line-data.tsv");
using (var env = EnvHelper.NewTestEnvironment(seed: 1, conc: 1))
{
// Pipeline
var loader = new TextLoader(env, args).Read(new MultiFileSource(trainFilename));
var trans = TextFeaturizingEstimator.Create(env, args2, loader);
// Train
var trainer = new SdcaBinaryTrainer(env, new SdcaBinaryTrainer.Arguments
{
NumThreads = 1
});
var cached = new CacheDataView(env, trans, prefetch: null);
var predictor = trainer.Fit(cached);
var scoreRoles = new RoleMappedData(trans, label: "Label", feature: "Features");
var trainRoles = new RoleMappedData(cached, label: "Label", feature: "Features");
return(ScoreUtils.GetScorer(predictor.Model, scoreRoles, env, trainRoles.Schema));
}
}
public void OVAUncalibrated()
{
var(pipeline, data) = GetMultiClassPipeline();
var sdcaTrainer = new SdcaBinaryTrainer(Env, "Label", "Features", advancedSettings: (s) => { s.MaxIterations = 100; s.Shuffle = true; s.NumThreads = 1; s.Calibrator = null; });
pipeline.Append(new Ova(Env, sdcaTrainer, useProbabilities: false))
.Append(new KeyToValueMappingEstimator(Env, "PredictedLabel"));
TestEstimatorCore(pipeline, data);
Done();
}
private IDataScorerTransform _TrainSentiment()
{
bool normalize = true;
var args = new TextLoader.Options()
{
Separators = new[] { '\t' },
HasHeader = true,
Columns = new[] {
new TextLoader.Column("Label", DataKind.Boolean, 0),
new TextLoader.Column("SentimentText", DataKind.String, 1)
}
};
var args2 = new TextFeaturizingEstimator.Options()
{
KeepDiacritics = false,
KeepPunctuations = false,
CaseMode = TextNormalizingEstimator.CaseMode.Lower,
OutputTokens = true,
UsePredefinedStopWordRemover = true,
Norm = normalize ? TextFeaturizingEstimator.NormFunction.L2 : TextFeaturizingEstimator.NormFunction.None,
CharFeatureExtractor = new WordBagEstimator.Options()
{
NgramLength = 3, UseAllLengths = false
},
WordFeatureExtractor = new WordBagEstimator.Options()
{
NgramLength = 2, UseAllLengths = true
},
};
var trainFilename = FileHelper.GetTestFile("wikipedia-detox-250-line-data.tsv");
/*using (*/
var env = EnvHelper.NewTestEnvironment(seed: 1, conc: 1);
{
// Pipeline
var loader = new TextLoader(env, args).Load(new MultiFileSource(trainFilename));
var trans = TextFeaturizingEstimator.Create(env, args2, loader);
// Train
var trainer = new SdcaBinaryTrainer(env, new SdcaBinaryTrainer.Options
{
});
var cached = new CacheDataView(env, trans, prefetch: null);
var predictor = trainer.Fit(cached);
var scoreRoles = new RoleMappedData(trans, label: "Label", feature: "Features");
var trainRoles = new RoleMappedData(cached, label: "Label", feature: "Features");
return(ScoreUtils.GetScorer(predictor.Model, scoreRoles, env, trainRoles.Schema));
}
}
public void Pkpd()
{
var(pipeline, data) = GetMultiClassPipeline();
var sdcaTrainer = new SdcaBinaryTrainer(Env, "Label", "Features", advancedSettings: (s) => { s.MaxIterations = 100; s.Shuffle = true; s.NumThreads = 1; });
pipeline.Append(new Pkpd(Env, sdcaTrainer))
.Append(new KeyToValueMappingEstimator(Env, "PredictedLabel"));
TestEstimatorCore(pipeline, data);
Done();
}
/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and a custom loss.
/// Note that because we cannot be sure that all loss functions will produce naturally calibrated outputs, setting
/// a custom loss function will not produce a calibrated probability column.
/// </summary>
/// <param name="ctx">The binary classification context trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="loss">The custom loss.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="l2Const">The L2 regularization hyperparameter.</param>
/// <param name="l1Threshold">The L1 regularization hyperparameter. Higher values will tend to lead to more sparse model.</param>
/// <param name="maxIterations">The maximum number of passes to perform over the data.</param>
/// <param name="advancedSettings">A delegate to set more settings.
/// The settings here will override the ones provided in the direct method signature,
/// if both are present and have different values.
/// The columns names, however need to be provided directly, not through the <paramref name="advancedSettings"/>.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), and the predicted label.</returns>
/// <seealso cref="Sdca(BinaryClassificationContext.BinaryClassificationTrainers, Scalar{bool}, Vector{float}, Scalar{float}, float?, float?, int?, Action{SdcaBinaryTrainer.Arguments}, Action{LinearBinaryModelParameters, ParameterMixingCalibratedPredictor})"/>
public static (Scalar <float> score, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationContext.BinaryClassificationTrainers ctx,
Scalar <bool> label, Vector <float> features,
ISupportSdcaClassificationLoss loss,
Scalar <float> weights = null,
float?l2Const = null,
float?l1Threshold = null,
int?maxIterations = null,
Action <SdcaBinaryTrainer.Arguments> advancedSettings = null,
Action <LinearBinaryModelParameters> onFit = null
)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValue(loss, nameof(loss));
Contracts.CheckValueOrNull(weights);
Contracts.CheckParam(!(l2Const < 0), nameof(l2Const), "Must not be negative, if specified.");
Contracts.CheckParam(!(l1Threshold < 0), nameof(l1Threshold), "Must not be negative, if specified.");
Contracts.CheckParam(!(maxIterations < 1), nameof(maxIterations), "Must be positive if specified");
Contracts.CheckValueOrNull(advancedSettings);
Contracts.CheckValueOrNull(onFit);
bool hasProbs = loss is LogLoss;
var rec = new TrainerEstimatorReconciler.BinaryClassifierNoCalibration(
(env, labelName, featuresName, weightsName) =>
{
var trainer = new SdcaBinaryTrainer(env, labelName, featuresName, weightsName, loss, l2Const, l1Threshold, maxIterations, advancedSettings);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
var model = trans.Model;
if (model is ParameterMixingCalibratedPredictor cali)
{
onFit((LinearBinaryModelParameters)cali.SubPredictor);
}
else
{
onFit((LinearBinaryModelParameters)model);
}
}));
}
return(trainer);
}, label, features, weights, hasProbs);
return(rec.Output);
}
public void MetacomponentsFeaturesRenamed()
{
var data = new TextLoader(Env, TestDatasets.irisData.GetLoaderColumns(), separatorChar: ',')
.Read(GetDataPath(TestDatasets.irisData.trainFilename));
var sdcaTrainer = new SdcaBinaryTrainer(Env, "Label", "Vars", advancedSettings: (s) => { s.MaxIterations = 100; s.Shuffle = true; s.NumThreads = 1; });
var pipeline = new ColumnConcatenatingEstimator(Env, "Vars", "SepalLength", "SepalWidth", "PetalLength", "PetalWidth")
.Append(new ValueToKeyMappingEstimator(Env, "Label"), TransformerScope.TrainTest)
.Append(new Ova(Env, sdcaTrainer))
.Append(new KeyToValueMappingEstimator(Env, "PredictedLabel"));
var model = pipeline.Fit(data);
TestEstimatorCore(pipeline, data);
Done();
}
public void SdcaWorkout()
{
var dataPath = GetDataPath("breast-cancer.txt");
var data = TextLoader.CreateReader(Env, ctx => (Label: ctx.LoadFloat(0), Features: ctx.LoadFloat(1, 10)))
.Read(dataPath);
IEstimator <ITransformer> est = new SdcaBinaryTrainer(Env, "Features", "Label", advancedSettings: (s) => s.ConvergenceTolerance = 1e-2f);
TestEstimatorCore(est, data.AsDynamic);
est = new SdcaRegressionTrainer(Env, "Features", "Label", advancedSettings: (s) => s.ConvergenceTolerance = 1e-2f);
TestEstimatorCore(est, data.AsDynamic);
est = new SdcaMultiClassTrainer(Env, "Features", "Label", advancedSettings: (s) => s.ConvergenceTolerance = 1e-2f);
TestEstimatorCore(est, data.AsDynamic);
Done();
}
/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and a custom loss.
/// Note that because we cannot be sure that all loss functions will produce naturally calibrated outputs, setting
/// a custom loss function will not produce a calibrated probability column.
/// </summary>
/// <param name="ctx">The binary classification context trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="loss">The custom loss.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="options">Advanced arguments to the algorithm.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), and the predicted label.</returns>
public static (Scalar <float> score, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationContext.BinaryClassificationTrainers ctx,
Scalar <bool> label, Vector <float> features,
Scalar <float> weights,
ISupportSdcaClassificationLoss loss,
SdcaBinaryTrainer.Options options,
Action <LinearBinaryModelParameters> onFit = null
)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValueOrNull(weights);
Contracts.CheckValueOrNull(options);
Contracts.CheckValueOrNull(onFit);
bool hasProbs = loss is LogLoss;
var rec = new TrainerEstimatorReconciler.BinaryClassifierNoCalibration(
(env, labelName, featuresName, weightsName) =>
{
options.FeatureColumn = featuresName;
options.LabelColumn = labelName;
var trainer = new SdcaBinaryTrainer(env, options);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
var model = trans.Model;
if (model is ParameterMixingCalibratedPredictor cali)
{
onFit((LinearBinaryModelParameters)cali.SubPredictor);
}
else
{
onFit((LinearBinaryModelParameters)model);
}
}));
}
return(trainer);
}, label, features, weights, hasProbs);
return(rec.Output);
}
/// <summary>
/// Predict a target using a linear binary classification model trained with the SDCA trainer, and log-loss.
/// </summary>
/// <param name="ctx">The binary classification context trainer object.</param>
/// <param name="label">The label, or dependent variable.</param>
/// <param name="features">The features, or independent variables.</param>
/// <param name="weights">The optional example weights.</param>
/// <param name="l2Const">The L2 regularization hyperparameter.</param>
/// <param name="l1Threshold">The L1 regularization hyperparameter. Higher values will tend to lead to more sparse model.</param>
/// <param name="maxIterations">The maximum number of passes to perform over the data.</param>
/// <param name="advancedSettings">A delegate to set more settings.
/// The settings here will override the ones provided in the direct method signature,
/// if both are present and have different values.
/// The columns names, however need to be provided directly, not through the <paramref name="advancedSettings"/>.</param>
/// <param name="onFit">A delegate that is called every time the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}.Fit(DataView{TInShape})"/> method is called on the
/// <see cref="Estimator{TInShape, TOutShape, TTransformer}"/> instance created out of this. This delegate will receive
/// the linear model that was trained, as well as the calibrator on top of that model. Note that this action cannot change the
/// result in any way; it is only a way for the caller to be informed about what was learnt.</param>
/// <returns>The set of output columns including in order the predicted binary classification score (which will range
/// from negative to positive infinity), the calibrated prediction (from 0 to 1), and the predicted label.</returns>
/// <example>
/// <format type="text/markdown">
/// <]
/// ]]></format>
/// </example>
public static (Scalar <float> score, Scalar <float> probability, Scalar <bool> predictedLabel) Sdca(
this BinaryClassificationContext.BinaryClassificationTrainers ctx,
Scalar <bool> label, Vector <float> features, Scalar <float> weights = null,
float?l2Const = null,
float?l1Threshold = null,
int?maxIterations = null,
Action <SdcaBinaryTrainer.Arguments> advancedSettings = null,
Action <LinearBinaryModelParameters, ParameterMixingCalibratedPredictor> onFit = null)
{
Contracts.CheckValue(label, nameof(label));
Contracts.CheckValue(features, nameof(features));
Contracts.CheckValueOrNull(weights);
Contracts.CheckParam(!(l2Const < 0), nameof(l2Const), "Must not be negative, if specified.");
Contracts.CheckParam(!(l1Threshold < 0), nameof(l1Threshold), "Must not be negative, if specified.");
Contracts.CheckParam(!(maxIterations < 1), nameof(maxIterations), "Must be positive if specified");
Contracts.CheckValueOrNull(advancedSettings);
Contracts.CheckValueOrNull(onFit);
var rec = new TrainerEstimatorReconciler.BinaryClassifier(
(env, labelName, featuresName, weightsName) =>
{
var trainer = new SdcaBinaryTrainer(env, labelName, featuresName, weightsName, loss: new LogLoss(), l2Const, l1Threshold, maxIterations, advancedSettings);
if (onFit != null)
{
return(trainer.WithOnFitDelegate(trans =>
{
// Under the default log-loss we assume a calibrated predictor.
var model = trans.Model;
var cali = (ParameterMixingCalibratedPredictor)model;
var pred = (LinearBinaryModelParameters)cali.SubPredictor;
onFit(pred, cali);
}));
}
return(trainer);
}, label, features, weights);
return(rec.Output);
}
