DataOperationsCatalog.CrossValidationSplit Metodo
Definizione
Importante
Alcune informazioni sono relative alla release non definitiva del prodotto, che potrebbe subire modifiche significative prima della release definitiva. Microsoft non riconosce alcuna garanzia, espressa o implicita, in merito alle informazioni qui fornite.
Suddividere il set di dati in piega di convalida incrociata del set di training e del set di test.
Rispetta l'oggetto samplingKeyColumnName se specificato.
public System.Collections.Generic.IReadOnlyList<Microsoft.ML.DataOperationsCatalog.TrainTestData> CrossValidationSplit (Microsoft.ML.IDataView data, int numberOfFolds = 5, string samplingKeyColumnName = default, int? seed = default);member this.CrossValidationSplit : Microsoft.ML.IDataView * int * string * Nullable<int> -> System.Collections.Generic.IReadOnlyList<Microsoft.ML.DataOperationsCatalog.TrainTestData>Public Function CrossValidationSplit (data As IDataView, Optional numberOfFolds As Integer = 5, Optional samplingKeyColumnName As String = Nothing, Optional seed As Nullable(Of Integer) = Nothing) As IReadOnlyList(Of DataOperationsCatalog.TrainTestData)Parametri
- data
- IDataView
Set di dati da dividere.
- numberOfFolds
- Int32
Numero di volte di convalida incrociata.
- samplingKeyColumnName
- String
Nome di una colonna da usare per raggruppare le righe. Se due esempi condividono lo stesso valore di samplingKeyColumnName, sono garantiti che vengano visualizzati nello stesso subset (training o test). Questa operazione può essere usata per garantire che nessuna perdita di etichette dal training al set di test.
Si noti che quando si esegue un esperimento di classificazione, deve samplingKeyColumnName essere la colonna GroupId.
Se null non verrà eseguito alcun raggruppamento di righe.
Inizializzazione per il generatore di numeri casuali usato per selezionare le righe per le piegazioni di convalida incrociata.
Restituisce
Esempio
using System;
using System.Collections.Generic;
using Microsoft.ML;
namespace Samples.Dynamic
{
/// <summary>
/// Sample class showing how to use CrossValidationSplit.
/// </summary>
public static class CrossValidationSplit
{
public static void Example()
{
// Creating the ML.Net IHostEnvironment object, needed for the pipeline.
var mlContext = new MLContext();
// Generate some data points.
var examples = GenerateRandomDataPoints(10);
// Convert the examples list to an IDataView object, which is consumable
// by ML.NET API.
var dataview = mlContext.Data.LoadFromEnumerable(examples);
// Cross validation splits your data randomly into set of "folds", and
// creates groups of Train and Test sets, where for each group, one fold
// is the Test and the rest of the folds the Train. So below, we specify
// Group column as the column containing the sampling keys. If we pass
// that column to cross validation it would be used to break data into
// certain chunks.
var folds = mlContext.Data
.CrossValidationSplit(dataview, numberOfFolds: 3,
samplingKeyColumnName: "Group");
var trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[0].TrainSet,
reuseRowObject: false);
var testSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[0].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 1], [Features, 0.8173254]
// [Group, 2], [Features, 0.7680227]
// [Group, 1], [Features, 0.2060332]
// [Group, 2], [Features, 0.5588848]
// [Group, 1], [Features, 0.4421779]
// [Group, 2], [Features, 0.9775497]
//
// The data in the Test split.
// [Group, 0], [Features, 0.7262433]
// [Group, 0], [Features, 0.5581612]
// [Group, 0], [Features, 0.9060271]
// [Group, 0], [Features, 0.2737045]
trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[1].TrainSet,
reuseRowObject: false);
testSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[1].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 0], [Features, 0.7262433]
// [Group, 2], [Features, 0.7680227]
// [Group, 0], [Features, 0.5581612]
// [Group, 2], [Features, 0.5588848]
// [Group, 0], [Features, 0.9060271]
// [Group, 2], [Features, 0.9775497]
// [Group, 0], [Features, 0.2737045]
//
// The data in the Test split.
// [Group, 1], [Features, 0.8173254]
// [Group, 1], [Features, 0.2060332]
// [Group, 1], [Features, 0.4421779]
trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[2].TrainSet,
reuseRowObject: false);
testSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[2].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 0], [Features, 0.7262433]
// [Group, 1], [Features, 0.8173254]
// [Group, 0], [Features, 0.5581612]
// [Group, 1], [Features, 0.2060332]
// [Group, 0], [Features, 0.9060271]
// [Group, 1], [Features, 0.4421779]
// [Group, 0], [Features, 0.2737045]
//
// The data in the Test split.
// [Group, 2], [Features, 0.7680227]
// [Group, 2], [Features, 0.5588848]
// [Group, 2], [Features, 0.9775497]
// Example of a split without specifying a sampling key column.
folds = mlContext.Data.CrossValidationSplit(dataview, numberOfFolds: 3);
trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[0].TrainSet,
reuseRowObject: false);
testSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[0].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 0], [Features, 0.7262433]
// [Group, 1], [Features, 0.8173254]
// [Group, 2], [Features, 0.7680227]
// [Group, 0], [Features, 0.5581612]
// [Group, 1], [Features, 0.2060332]
// [Group, 1], [Features, 0.4421779]
// [Group, 2], [Features, 0.9775497]
// [Group, 0], [Features, 0.2737045]
//
// The data in the Test split.
// [Group, 2], [Features, 0.5588848]
// [Group, 0], [Features, 0.9060271]
trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[1].TrainSet,
reuseRowObject: false);
testSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[1].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 2], [Features, 0.7680227]
// [Group, 0], [Features, 0.5581612]
// [Group, 1], [Features, 0.2060332]
// [Group, 2], [Features, 0.5588848]
// [Group, 0], [Features, 0.9060271]
// [Group, 1], [Features, 0.4421779]
//
// The data in the Test split.
// [Group, 0], [Features, 0.7262433]
// [Group, 1], [Features, 0.8173254]
// [Group, 2], [Features, 0.9775497]
// [Group, 0], [Features, 0.2737045]
trainSet = mlContext.Data
.CreateEnumerable<DataPoint>(folds[2].TrainSet,
reuseRowObject: false);
testSet = mlContext.Data.CreateEnumerable<DataPoint>(folds[2].TestSet,
reuseRowObject: false);
PrintPreviewRows(trainSet, testSet);
// The data in the Train split.
// [Group, 0], [Features, 0.7262433]
// [Group, 1], [Features, 0.8173254]
// [Group, 2], [Features, 0.5588848]
// [Group, 0], [Features, 0.9060271]
// [Group, 2], [Features, 0.9775497]
// [Group, 0], [Features, 0.2737045]
//
// The data in the Test split.
// [Group, 2], [Features, 0.7680227]
// [Group, 0], [Features, 0.5581612]
// [Group, 1], [Features, 0.2060332]
// [Group, 1], [Features, 0.4421779]
}
private static IEnumerable<DataPoint> GenerateRandomDataPoints(int count,
int seed = 0)
{
var random = new Random(seed);
for (int i = 0; i < count; i++)
{
yield return new DataPoint
{
Group = i % 3,
// Create random features that are correlated with label.
Features = (float)random.NextDouble()
};
}
}
// Example with features and group column. A data set is a collection of
// such examples.
private class DataPoint
{
public float Group { get; set; }
public float Features { get; set; }
}
// print helper
private static void PrintPreviewRows(IEnumerable<DataPoint> trainSet,
IEnumerable<DataPoint> testSet)
{
Console.WriteLine($"The data in the Train split.");
foreach (var row in trainSet)
Console.WriteLine($"{row.Group}, {row.Features}");
Console.WriteLine($"\nThe data in the Test split.");
foreach (var row in testSet)
Console.WriteLine($"{row.Group}, {row.Features}");
}
}
}