/// <summary>
/// Initialize (or reinitialize) the stopwords remover. Also called by the constructor.
/// </summary>
/// <param name="CustomStopWords"></param>
public void InitializeStopWordsRemover(string[] CustomStopWords = null)
{
_mlContext = new MLContext();
_emptySamplesList = new List <TextData>();
_emptyDataView = _mlContext.Data.LoadFromEnumerable(_emptySamplesList);
// stop words
_stopWordsTextPipeline = _mlContext.Transforms.Text.TokenizeIntoWords("Words", "Text")
.Append(_mlContext.Transforms.Text.RemoveDefaultStopWords("WordsWithoutStopWords", "Words", language: StopWordsRemovingEstimator.Language.English));
if (CustomStopWords == null)
{
_textPipeline = _stopWordsTextPipeline.Fit(_emptyDataView);
_mode = removeStopWordswMode.Default;
}
else
{
var textPipeline = _mlContext.Transforms.Text.TokenizeIntoWords("Words",
"Text")
.Append(_mlContext.Transforms.Text.RemoveStopWords(
"WordsWithoutStopWords", "Words", stopwords:
CustomStopWords));
_mode = removeStopWordswMode.Custom;
}
_predictionEngine = _mlContext.Model.CreatePredictionEngine <TextData, TransformedTextData>(_textPipeline);
}
public void Train(string trainingFileName)
{
//1. First, we check to make sure that the training data filename exists
if (!File.Exists(trainingFileName))
{
Console.WriteLine($"Failed to find training data file ({trainingFileName}");
return;
}
//2. Use the LoadFromTextFile helper method that ML.NET provides to assist with
//the loading of text files into an IDataView object
IDataView trainingDataView = MlContext.Data.LoadFromTextFile <RestaurantFeedback>(trainingFileName);
//3. Use the TrainTestSplit method that ML.NET provides to create a test set from the
//main training data
DataOperationsCatalog.TrainTestData dataSplit = MlContext.Data.TrainTestSplit(trainingDataView, testFraction: 0.2);
//4. Firstly, we create the pipeline
Microsoft.ML.Transforms.Text.TextFeaturizingEstimator dataProcessPipeline = MlContext.Transforms.Text.FeaturizeText(
outputColumnName: "Features",
inputColumnName: nameof(RestaurantFeedback.Text));
//5. Next, we instantiate our Trainer class
Microsoft.ML.Trainers.SdcaLogisticRegressionBinaryTrainer sdcaRegressionTrainer = MlContext.BinaryClassification.Trainers.SdcaLogisticRegression(
labelColumnName: nameof(RestaurantFeedback.Label),
featureColumnName: "Features");
//6. Then, we complete the pipeline by appending the trainer we instantiated
//previously
Microsoft.ML.Data.EstimatorChain <Microsoft.ML.Data.BinaryPredictionTransformer <Microsoft.ML.Calibrators.CalibratedModelParametersBase <Microsoft.ML.Trainers.LinearBinaryModelParameters, Microsoft.ML.Calibrators.PlattCalibrator> > > trainingPipeline = dataProcessPipeline.Append(sdcaRegressionTrainer);
//7. Next, we train the model with the dataset we created earlier in the chapter
ITransformer trainedModel = trainingPipeline.Fit(dataSplit.TrainSet);
MlContext.Model.Save(trainedModel, dataSplit.TrainSet.Schema, ModelPath);
//8. We save our newly created model to the filename specified, matching the
//training set's schema
IDataView testSetTransform = trainedModel.Transform(dataSplit.TestSet);
//9. Now, we transform our newly created model with the test set we created earlier
Microsoft.ML.Data.CalibratedBinaryClassificationMetrics modelMetrics = MlContext.BinaryClassification.Evaluate(
data: testSetTransform,
labelColumnName: nameof(RestaurantFeedback.Label),
scoreColumnName: nameof(RestaurantPrediction.Score));
//10. Finally, we will use the testSetTransform function created previously and
//pass it into the BinaryClassification class's Evaluate method
//This method allows us to generate model metrics. We then print the main metrics using the
//trained model with the test set. We will dive into these properties specifically in the
//Evaluating the Model section of this chapter.
Console.WriteLine($"Area Under Curve: {modelMetrics.AreaUnderRocCurve:P2}{Environment.NewLine}" +
$"Area Under Precision Recall Curve: {modelMetrics.AreaUnderPrecisionRecallCurve:P2}{Environment.NewLine}" +
$"Accuracy: {modelMetrics.Accuracy:P2}{Environment.NewLine}" +
$"F1Score: {modelMetrics.F1Score:P2}{Environment.NewLine}" +
$"Positive Recall: {modelMetrics.PositiveRecall:#.##}{Environment.NewLine}" +
$"Negative Recall: {modelMetrics.NegativeRecall:#.##}{Environment.NewLine}");
}
/// <summary>
/// Train/create the ML model.
/// </summary>
/// <param name="trainDataView"></param>
/// <param name="pipeline"></param>
/// <returns>Trained Model</returns>
private static ITransformer GetTrainedModel(IDataView trainDataView, Microsoft.ML.Data.EstimatorChain <KeyToValueMappingTransformer> pipeline)
{
Console.WriteLine("*** Training the image classification model with DNN Transfer Learning on top of the selected pre-trained model/architecture ***");
var watch = Stopwatch.StartNew();
ITransformer trainedModel = pipeline.Fit(trainDataView);
watch.Stop();
var elapsedMs = watch.ElapsedMilliseconds;
Console.WriteLine($"Training with transfer learning took: {elapsedMs / 1000} seconds");
return(trainedModel);
}
public void Train(string trainingFileName)
{
if (!File.Exists(trainingFileName))
{
Console.WriteLine($"Failed to find training data file ({trainingFileName}");
return;
}
//The first change is the use of a comma to separate the data
IDataView trainingDataView = MlContext.Data.LoadFromTextFile <EmploymentHistory>(trainingFileName, ',');
DataOperationsCatalog.TrainTestData dataSplit = MlContext.Data.TrainTestSplit(trainingDataView, testFraction: 0.4);
Microsoft.ML.Data.EstimatorChain <Microsoft.ML.Data.TransformerChain <Microsoft.ML.Data.ColumnConcatenatingTransformer> > dataProcessPipeline = MlContext.Transforms.CopyColumns("Label", nameof(EmploymentHistory.DurationInMonths))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.IsMarried)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.BSDegree)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.MSDegree)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.YearsExperience))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.AgeAtHire)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.HasKids)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.WithinMonthOfVesting)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.DeskDecorations)))
.Append(MlContext.Transforms.NormalizeMeanVariance(nameof(EmploymentHistory.LongCommute)))
.Append(MlContext.Transforms.Concatenate("Features",
typeof(EmploymentHistory).ToPropertyList <EmploymentHistory>(nameof(EmploymentHistory.DurationInMonths)))));
//We can then create the Sdca trainer using the default parameters
Microsoft.ML.Trainers.SdcaRegressionTrainer trainer = MlContext.Regression.Trainers.Sdca(labelColumnName: "Label", featureColumnName: "Features");
Microsoft.ML.Data.EstimatorChain <Microsoft.ML.Data.RegressionPredictionTransformer <Microsoft.ML.Trainers.LinearRegressionModelParameters> > trainingPipeline = dataProcessPipeline.Append(trainer);
ITransformer trainedModel = trainingPipeline.Fit(dataSplit.TrainSet);
MlContext.Model.Save(trainedModel, dataSplit.TrainSet.Schema, ModelPath);
IDataView testSetTransform = trainedModel.Transform(dataSplit.TestSet);
//Lastly, we call the Regression.Evaluate method to provide regression specific metrics
Microsoft.ML.Data.RegressionMetrics modelMetrics = MlContext.Regression.Evaluate(testSetTransform);
Console.WriteLine($"Loss Function: {modelMetrics.LossFunction:0.##}{Environment.NewLine}" +
$"Mean Absolute Error: {modelMetrics.MeanAbsoluteError:#.##}{Environment.NewLine}" +
$"Mean Squared Error: {modelMetrics.MeanSquaredError:#.##}{Environment.NewLine}" +
$"RSquared: {modelMetrics.RSquared:0.##}{Environment.NewLine}" +
$"Root Mean Squared Error: {modelMetrics.RootMeanSquaredError:#.##}");
}