StandardTrainersCatalog.SgdNonCalibrated Metodo

  • Riferimento

Definizione

Spazio dei nomi:
Microsoft.ML
Assembly:
Microsoft.ML.StandardTrainers.dll
Pacchetto:
Microsoft.ML v3.0.1
Pacchetto:
Microsoft.ML v1.0.0
Pacchetto:
Microsoft.ML v1.1.0
Pacchetto:
Microsoft.ML v1.2.0
Pacchetto:
Microsoft.ML v1.3.1
Pacchetto:
Microsoft.ML v1.4.0
Pacchetto:
Microsoft.ML v1.5.5
Pacchetto:
Microsoft.ML v1.6.0
Pacchetto:
Microsoft.ML v1.7.0
Pacchetto:
Microsoft.ML v2.0.0

Overload

SgdNonCalibrated(BinaryClassificationCatalog+BinaryClassificationTrainers, SgdNonCalibratedTrainer+Options)

Creare SgdNonCalibratedTrainer con opzioni avanzate, che stimano una destinazione usando un modello di classificazione lineare. La discesa del gradiente stocastico (SGD) è un algoritmo iterativo che ottimizza una funzione obiettivo differenziabile.
SgdNonCalibrated(BinaryClassificationCatalog+BinaryClassificationTrainers, String, String, String, IClassificationLoss, Int32, Double, Single)

Creare SgdNonCalibratedTrainer, che stima una destinazione usando un modello di classificazione lineare. La discesa del gradiente stocastico (SGD) è un algoritmo iterativo che ottimizza una funzione obiettivo differenziabile.

SgdNonCalibrated(BinaryClassificationCatalog+BinaryClassificationTrainers, SgdNonCalibratedTrainer+Options)

Creare SgdNonCalibratedTrainer con opzioni avanzate, che stimano una destinazione usando un modello di classificazione lineare. La discesa del gradiente stocastico (SGD) è un algoritmo iterativo che ottimizza una funzione obiettivo differenziabile.

public static Microsoft.ML.Trainers.SgdNonCalibratedTrainer SgdNonCalibrated (this Microsoft.ML.BinaryClassificationCatalog.BinaryClassificationTrainers catalog, Microsoft.ML.Trainers.SgdNonCalibratedTrainer.Options options);
static member SgdNonCalibrated : Microsoft.ML.BinaryClassificationCatalog.BinaryClassificationTrainers * Microsoft.ML.Trainers.SgdNonCalibratedTrainer.Options -> Microsoft.ML.Trainers.SgdNonCalibratedTrainer
<Extension()>
Public Function SgdNonCalibrated (catalog As BinaryClassificationCatalog.BinaryClassificationTrainers, options As SgdNonCalibratedTrainer.Options) As SgdNonCalibratedTrainer

Parametri

catalog
BinaryClassificationCatalog.BinaryClassificationTrainers

Oggetto di training del catalogo di classificazione binaria.

options
SgdNonCalibratedTrainer.Options

Opzioni del formatore.

Restituisce

SgdNonCalibratedTrainer

Esempio

using System;
using System.Collections.Generic;
using System.Linq;
using Microsoft.ML;
using Microsoft.ML.Data;
using Microsoft.ML.Trainers;

namespace Samples.Dynamic.Trainers.BinaryClassification
{
    public static class SgdNonCalibratedWithOptions
    {
        public static void Example()
        {
            // Create a new context for ML.NET operations. It can be used for
            // exception tracking and logging, as a catalog of available operations
            // and as the source of randomness. Setting the seed to a fixed number
            // in this example to make outputs deterministic.
            var mlContext = new MLContext(seed: 0);

            // Create a list of training data points.
            var dataPoints = GenerateRandomDataPoints(1000);

            // Convert the list of data points to an IDataView object, which is
            // consumable by ML.NET API.
            var trainingData = mlContext.Data.LoadFromEnumerable(dataPoints);

            // Define trainer options.
            var options = new SgdNonCalibratedTrainer.Options
            {
                LearningRate = 0.01,
                NumberOfIterations = 10,
                L2Regularization = 1e-7f
            };

            // Define the trainer.
            var pipeline = mlContext.BinaryClassification.Trainers
                .SgdNonCalibrated(options);

            // Train the model.
            var model = pipeline.Fit(trainingData);

            // Create testing data. Use different random seed to make it different
            // from training data.
            var testData = mlContext.Data
                .LoadFromEnumerable(GenerateRandomDataPoints(500, seed: 123));

            // Run the model on test data set.
            var transformedTestData = model.Transform(testData);

            // Convert IDataView object to a list.
            var predictions = mlContext.Data
                .CreateEnumerable<Prediction>(transformedTestData,
                reuseRowObject: false).ToList();

            // Print 5 predictions.
            foreach (var p in predictions.Take(5))
                Console.WriteLine($"Label: {p.Label}, "
                    + $"Prediction: {p.PredictedLabel}");

            // Expected output:
            //   Label: True, Prediction: False
            //   Label: False, Prediction: False
            //   Label: True, Prediction: True
            //   Label: True, Prediction: True
            //   Label: False, Prediction: False

            // Evaluate the overall metrics.
            var metrics = mlContext.BinaryClassification
                .EvaluateNonCalibrated(transformedTestData);

            PrintMetrics(metrics);

            // Expected output:
            //   Accuracy: 0.59
            //   AUC: 0.61
            //   F1 Score: 0.41
            //   Negative Precision: 0.57
            //   Negative Recall: 0.85
            //   Positive Precision: 0.64
            //   Positive Recall: 0.30
            //
            //   TEST POSITIVE RATIO:    0.4760 (238.0/(238.0+262.0))
            //   Confusion table
            //             ||======================
            //   PREDICTED || positive | negative | Recall
            //   TRUTH     ||======================
            //    positive ||      137 |      101 | 0.5756
            //    negative ||      118 |      144 | 0.5496
            //             ||======================
            //   Precision ||   0.5373 |   0.5878 |
        }

        private static IEnumerable<DataPoint> GenerateRandomDataPoints(int count,
            int seed = 0)

        {
            var random = new Random(seed);
            float randomFloat() => (float)random.NextDouble();
            for (int i = 0; i < count; i++)
            {
                var label = randomFloat() > 0.5f;
                yield return new DataPoint
                {
                    Label = label,
                    // Create random features that are correlated with the label.
                    // For data points with false label, the feature values are
                    // slightly increased by adding a constant.
                    Features = Enumerable.Repeat(label, 50)
                        .Select(x => x ? randomFloat() : randomFloat() +
                        0.03f).ToArray()

                };
            }
        }

        // Example with label and 50 feature values. A data set is a collection of
        // such examples.
        private class DataPoint
        {
            public bool Label { get; set; }
            [VectorType(50)]
            public float[] Features { get; set; }
        }

        // Class used to capture predictions.
        private class Prediction
        {
            // Original label.
            public bool Label { get; set; }
            // Predicted label from the trainer.
            public bool PredictedLabel { get; set; }
        }

        // Pretty-print BinaryClassificationMetrics objects.
        private static void PrintMetrics(BinaryClassificationMetrics metrics)
        {
            Console.WriteLine($"Accuracy: {metrics.Accuracy:F2}");
            Console.WriteLine($"AUC: {metrics.AreaUnderRocCurve:F2}");
            Console.WriteLine($"F1 Score: {metrics.F1Score:F2}");
            Console.WriteLine($"Negative Precision: " +
                $"{metrics.NegativePrecision:F2}");

            Console.WriteLine($"Negative Recall: {metrics.NegativeRecall:F2}");
            Console.WriteLine($"Positive Precision: " +
                $"{metrics.PositivePrecision:F2}");

            Console.WriteLine($"Positive Recall: {metrics.PositiveRecall:F2}\n");
            Console.WriteLine(metrics.ConfusionMatrix.GetFormattedConfusionTable());
        }
    }
}

Si applica a

SgdNonCalibrated(BinaryClassificationCatalog+BinaryClassificationTrainers, String, String, String, IClassificationLoss, Int32, Double, Single)

Creare SgdNonCalibratedTrainer, che stima una destinazione usando un modello di classificazione lineare. La discesa del gradiente stocastico (SGD) è un algoritmo iterativo che ottimizza una funzione obiettivo differenziabile.

public static Microsoft.ML.Trainers.SgdNonCalibratedTrainer SgdNonCalibrated (this Microsoft.ML.BinaryClassificationCatalog.BinaryClassificationTrainers catalog, string labelColumnName = "Label", string featureColumnName = "Features", string exampleWeightColumnName = default, Microsoft.ML.Trainers.IClassificationLoss lossFunction = default, int numberOfIterations = 20, double learningRate = 0.01, float l2Regularization = 1E-06);
static member SgdNonCalibrated : Microsoft.ML.BinaryClassificationCatalog.BinaryClassificationTrainers * string * string * string * Microsoft.ML.Trainers.IClassificationLoss * int * double * single -> Microsoft.ML.Trainers.SgdNonCalibratedTrainer
<Extension()>
Public Function SgdNonCalibrated (catalog As BinaryClassificationCatalog.BinaryClassificationTrainers, Optional labelColumnName As String = "Label", Optional featureColumnName As String = "Features", Optional exampleWeightColumnName As String = Nothing, Optional lossFunction As IClassificationLoss = Nothing, Optional numberOfIterations As Integer = 20, Optional learningRate As Double = 0.01, Optional l2Regularization As Single = 1E-06) As SgdNonCalibratedTrainer

Parametri

catalog
BinaryClassificationCatalog.BinaryClassificationTrainers

Oggetto di training del catalogo di classificazione binaria.

labelColumnName
String

Nome della colonna dell'etichetta o variabile dipendente. I dati della colonna devono essere Boolean.

featureColumnName
String

Funzionalità o variabili indipendenti. I dati della colonna devono essere un vettore di dimensioni note di Single.

exampleWeightColumnName
String

Nome della colonna peso di esempio (facoltativo).

lossFunction
IClassificationLoss

Funzione di perdita ridotta al minimo nel processo di training. L'uso, ad esempio, HingeLoss porta a un formatore di macchine a vettori di supporto.

numberOfIterations
Int32

Il numero massimo di passaggi attraverso il set di dati di training; impostare su 1 per simulare l'apprendimento online.

learningRate
Double

Frequenza di apprendimento iniziale usata da SGD.

l2Regularization
Single

Peso L2 per la regolarizzazione.

Restituisce

SgdNonCalibratedTrainer

Esempio

using System;
using System.Collections.Generic;
using System.Linq;
using Microsoft.ML;
using Microsoft.ML.Data;

namespace Samples.Dynamic.Trainers.BinaryClassification
{
    public static class SgdNonCalibrated
    {
        public static void Example()
        {
            // Create a new context for ML.NET operations. It can be used for
            // exception tracking and logging, as a catalog of available operations
            // and as the source of randomness. Setting the seed to a fixed number
            // in this example to make outputs deterministic.
            var mlContext = new MLContext(seed: 0);

            // Create a list of training data points.
            var dataPoints = GenerateRandomDataPoints(1000);

            // Convert the list of data points to an IDataView object, which is
            // consumable by ML.NET API.
            var trainingData = mlContext.Data.LoadFromEnumerable(dataPoints);

            // Define the trainer.
            var pipeline = mlContext.BinaryClassification.Trainers
                .SgdNonCalibrated();

            // Train the model.
            var model = pipeline.Fit(trainingData);

            // Create testing data. Use different random seed to make it different
            // from training data.
            var testData = mlContext.Data
                .LoadFromEnumerable(GenerateRandomDataPoints(500, seed: 123));

            // Run the model on test data set.
            var transformedTestData = model.Transform(testData);

            // Convert IDataView object to a list.
            var predictions = mlContext.Data
                .CreateEnumerable<Prediction>(transformedTestData,
                reuseRowObject: false).ToList();

            // Print 5 predictions.
            foreach (var p in predictions.Take(5))
                Console.WriteLine($"Label: {p.Label}, "
                    + $"Prediction: {p.PredictedLabel}");

            // Expected output:
            //   Label: True, Prediction: False
            //   Label: False, Prediction: False
            //   Label: True, Prediction: True
            //   Label: True, Prediction: True
            //   Label: False, Prediction: False

            // Evaluate the overall metrics.
            var metrics = mlContext.BinaryClassification
                .EvaluateNonCalibrated(transformedTestData);

            PrintMetrics(metrics);

            // Expected output:
            //   Accuracy: 0.60
            //   AUC: 0.63
            //   F1 Score: 0.43
            //   Negative Precision: 0.58
            //   Negative Recall: 0.85
            //   Positive Precision: 0.66
            //   Positive Recall: 0.32
            //   
            //   TEST POSITIVE RATIO:    0.4760 (238.0/(238.0+262.0))
            //   Confusion table
            //             ||======================
            //   PREDICTED || positive | negative | Recall
            //   TRUTH     ||======================
            //    positive ||       76 |      162 | 0.3193
            //    negative ||       42 |      220 | 0.8397
            //             ||======================
            //   Precision ||   0.6441 |   0.5759 |
        }

        private static IEnumerable<DataPoint> GenerateRandomDataPoints(int count,
            int seed = 0)

        {
            var random = new Random(seed);
            float randomFloat() => (float)random.NextDouble();
            for (int i = 0; i < count; i++)
            {
                var label = randomFloat() > 0.5f;
                yield return new DataPoint
                {
                    Label = label,
                    // Create random features that are correlated with the label.
                    // For data points with false label, the feature values are
                    // slightly increased by adding a constant.
                    Features = Enumerable.Repeat(label, 50)
                        .Select(x => x ? randomFloat() : randomFloat() +
                        0.03f).ToArray()

                };
            }
        }

        // Example with label and 50 feature values. A data set is a collection of
        // such examples.
        private class DataPoint
        {
            public bool Label { get; set; }
            [VectorType(50)]
            public float[] Features { get; set; }
        }

        // Class used to capture predictions.
        private class Prediction
        {
            // Original label.
            public bool Label { get; set; }
            // Predicted label from the trainer.
            public bool PredictedLabel { get; set; }
        }

        // Pretty-print BinaryClassificationMetrics objects.
        private static void PrintMetrics(BinaryClassificationMetrics metrics)
        {
            Console.WriteLine($"Accuracy: {metrics.Accuracy:F2}");
            Console.WriteLine($"AUC: {metrics.AreaUnderRocCurve:F2}");
            Console.WriteLine($"F1 Score: {metrics.F1Score:F2}");
            Console.WriteLine($"Negative Precision: " +
                $"{metrics.NegativePrecision:F2}");

            Console.WriteLine($"Negative Recall: {metrics.NegativeRecall:F2}");
            Console.WriteLine($"Positive Precision: " +
                $"{metrics.PositivePrecision:F2}");

            Console.WriteLine($"Positive Recall: {metrics.PositiveRecall:F2}\n");
            Console.WriteLine(metrics.ConfusionMatrix.GetFormattedConfusionTable());
        }
    }
}

Si applica a