问ML.Net机器学习

EN

    private static string _appPath => Path.GetDirectoryName(Environment.GetCommandLineArgs()[0]);

    //TRAIN_DATA_FILEPATH: has the path to the dataset used to train the model.
    private static string TRAIN_DATA_FILEPATH => Path.Combine(_appPath, "..", "..", "..", "Data", "A.csv");
    //@"C:\Users\taqwa\Desktop\A.csv"
    private static string MODEL_FILEPATH = @"../../../../MyMLAppML.Model/MLModel.zip";

    // Create MLContext to be shared across the model creation workflow objects 
    // Set a random seed for repeatable/deterministic results across multiple trainings.
    private static MLContext mlContext = new MLContext(seed: 1);

    public static void CreateModel()
    {
        // Load Data
        //ModelInput is the input dataset class and has the following String fields: Cases, Algorith and InjuryOrIllness 
        IDataView trainingDataView = mlContext.Data.LoadFromTextFile<ModelInput>(
                                        path: TRAIN_DATA_FILEPATH,
                                        hasHeader: true,   //true if the Header property is not null; otherwise, false. The default is false.
                                        separatorChar: ',',
                                        allowQuoting: true,  //Whether the file can contain columns defined by a quoted string. Whether the input may include quoted values, which can contain separator characters, colons, and distinguish empty values from missing values. When true, consecutive separators denote a missing value and an empty value is denoted by "". When false, consecutive separators denote an empty value.
                                        allowSparse: false); //Whether the file can contain numerical vectors in sparse format.


        // Build training pipeline
        IEstimator<ITransformer> trainingPipeline = BuildTrainingPipeline(mlContext);

        // Evaluate quality of Model
    //    Evaluate(mlContext, trainingDataView, trainingPipeline);

        // Train Model
        ITransformer mlModel = TrainModel(mlContext, trainingDataView, trainingPipeline);

        // Save model
      //  SaveModel(mlContext, mlModel, MODEL_FILEPATH, trainingDataView.Schema);
    }

    public static IEstimator<ITransformer> BuildTrainingPipeline(MLContext mlContext)
    {
        // Data process configuration with pipeline data transformations 
        var dataProcessPipeline = mlContext.Transforms.Conversion.MapValueToKey("Algorithm", "Algorithm")
                                  //MapValueToKey: method to transform the Algorithm column into a numeric key type Algorithm column (a format accepted by classification algorithms) and add it as a new dataset column
                                  .Append(mlContext.Transforms.Categorical.OneHotEncoding(new[] { new InputOutputColumnPair("injuryOrIllness", "injuryOrIllness") }))
                                  //OneHotEncoding: which converts one or more input text columns specified in columns into as many columns of one-hot encoded vectors.
                                  .Append(mlContext.Transforms.Text.FeaturizeText("Cases_tf", "Cases"))
                                  //FeaturizeText which transforms the text (Cases_tf) columns into a numeric vector for each called Cases and Append the featurization to the pipeline
                                  .Append(mlContext.Transforms.Concatenate("Features", new[] { "injuryOrIllness", "Cases_tf" }))
                                  .Append(mlContext.Transforms.NormalizeMinMax("Features", "Features"))
                                  //AppendCacheCheckpoint to cache the DataView so when you iterate over the data multiple times using the cache might get better performance
                                  .AppendCacheCheckpoint(mlContext);


        // Set the training algorithm 
        //Here we used the AveragedPerceptron
        var trainer = mlContext.MulticlassClassification.Trainers.OneVersusAll(mlContext.BinaryClassification.Trainers.AveragedPerceptron(labelColumnName: "Algorithm", numberOfIterations: 10, featureColumnName: "Features"), labelColumnName: "Algorithm")
                                  .Append(mlContext.Transforms.Conversion.MapKeyToValue("PredictedLabel", "PredictedLabel"));
        //var trainer = mlContext.MulticlassClassification.Trainers.SdcaMaximumEntropy(labelColumnName: "Algorithm", featureColumnName: "Features")
        //              .Append(mlContext.Transforms.Conversion.MapKeyToValue("PredictedLabel", "PredictedLabel"));

        //OneVersusAllTrainer: which predicts a multiclass target using one-versus-all strategy with the binary classification estimator specified by binaryEstimator.
        var trainingPipeline = dataProcessPipeline.Append(trainer);


        return trainingPipeline;

    }


    public static ITransformer TrainModel(MLContext mlContext, IDataView trainingDataView, IEstimator<ITransformer> trainingPipeline)
    {
        Console.WriteLine("=============== Training  model ===============");

        //Fit(): method trains your model by transforming the dataset and applying the training. and return the trained model.
        ITransformer model = trainingPipeline.Fit(trainingDataView);
        Console.WriteLine($"{trainingDataView.Schema}");



        Console.WriteLine("=============== End of training process ===============");
        return model;
    }