This article demonstrates the process of creating, training, saving, and loading a spam detection AI model using ML.NET, but also emphasizes the reusability of the trained model. By following the steps in the article, you will be able to create a model that can be easily reused and integrated into your .NET applications, allowing you to effectively identify and filter out spam emails.

Prerequisites

  • Basic understanding of C#
  • Familiarity with ML.NET and machine learning concepts

Code Overview

    1. Import necessary namespaces:

      using System;
      using System.IO;
      using System.Linq;
      using Microsoft.ML;
      using Microsoft.ML.Data;
    
    1. Define the Email class and its properties:

      public class Email
      {
        public string Content { get; set; }
        public bool IsSpam { get; set; }
      }
    
    1. Create a sample dataset for training the model:

      var sampleData = new List<Email>
      {
        new Email { Content = "Buy cheap products now", IsSpam = true },
        new Email { Content = "Meeting at 3 PM", IsSpam = false },
      };
    
    1. Initialize a new MLContext, which is the main entry point to ML.NET:

      var mlContext = new MLContext();
    
    1. Load the sample data into an IDataView:

      var trainData = mlContext.Data.LoadFromEnumerable(sampleData);
    
    1. Define the data processing pipeline and the training algorithm (SdcaLogisticRegression):

      var pipeline = mlContext.Transforms.Text.FeaturizeText("Features", nameof(Email.Content))
        .Append(mlContext.BinaryClassification.Trainers.SdcaLogisticRegression());
    
    1. Train the model:

      var model = pipeline.Fit(trainData);
    
    1. Save the trained model as a .NET binary:

      mlContext.Model.Save(model, trainData.Schema, "model.zip");
    
    1. Load the saved model:

      var newMlContext = new MLContext();
      DataViewSchema modelSchema;
      ITransformer trainedModel = newMlContext.Model.Load("model.zip", out modelSchema);
    
    1. Create a prediction engine:

      var predictionEngine = mlContext.Model.CreatePredictionEngine<Email, SpamPrediction>(trainedModel);
    
    1. Test the model with a sample email:

      var sampleEmail = new Email { Content = "Special discount, buy now!" };
      var prediction = predictionEngine.Predict(sampleEmail);
    
    1. Output the prediction:

      Debug.WriteLine($"Email: '{sampleEmail.Content}' is {(prediction.IsSpam ? "spam" : "not spam")}");
    
    1. Assert that the prediction is correct:

      Assert.IsTrue(prediction.IsSpam);
    
    1. Verify that the model was saved:

      if(File.Exists("model.zip"))
        Assert.Pass();
      else
        Assert.Fail();
    

Conclusion

In this article, we explained a simple spam detection model in ML.NET and demonstrated how to train and test the model. This code can be extended to build more complex models, and can be used as a starting point for exploring machine learning in .NET.

Github Repo