ML.NET: Machine Learning for .NET Developers

Welcome, .NET developers! Explore the power of ML.NET, Microsoft's opensource machine learning framework. Leverage your existing C# skills to build intelligent applications. No Python or TensorFlow needed.

 Step by Step Project Implementation:

Create Console Application Project using C#

Download Micorsoft.ML Package

Step 1st:

using Microsoft.ML.Data;

namespace MLExample

{

    internal class StudentData

    {

        [LoadColumn(0)]

        public float StudyHours;

        [LoadColumn(1)]

        public float Attendance;

        [LoadColumn(2), ColumnName("Label")]

        public bool Passed;

    }

}

2) Create StudentPrediction Class

internal class StudentPrediction

{

    [ColumnName("PredictedLabel")]

    public bool Passed;

    public float Probability { get; set; }

    public float Score { get; set; }

}

3) Create student-data.csv file as a Data Source

5,90,True

2,60,False

8,95,True

1,50,False

6,85,True

3,65,False

7,92,True

2,40,False

4,70,True

1,30,False

4) Program.cs file

using Microsoft.ML;

using MLExample;

using System;

    class Program

    {

        static void Main(String[] args)

        {

        Console.WriteLine("Hello");

        var context = new MLContext();

        // 2. Load Data (update your file path here)

        var data = context.Data.LoadFromTextFile<StudentData>(

            path: @"d:\student-data.csv",

            hasHeader: false,

            separatorChar: ',');

        // 3. Define the pipeline

        var pipeline = context.Transforms

      .Concatenate("Features", nameof(StudentData.StudyHours), nameof(StudentData.Attendance))

      .Append(context.BinaryClassification.Trainers.SdcaLogisticRegression(

          new Microsoft.ML.Trainers.SdcaLogisticRegressionBinaryTrainer.Options

          {

              MaximumNumberOfIterations = 10

          }));

        Console.WriteLine("Training started...");

        // 4. Train the model

        var model = pipeline.Fit(data);

        Console.WriteLine("Training completed.");

        // 5. Create prediction engine (for single prediction)

        var predictor = context.Model.CreatePredictionEngine<StudentData, StudentPrediction>(model);

        // 6. Create sample input

        var newStudent = new StudentData

        {

            StudyHours = 4,

            Attendance = 80

        };

        // 7. Predict

        var result = predictor.Predict(newStudent);

        // 8. Output result

        Console.WriteLine($"Study Hours: {newStudent.StudyHours}, Attendance: {newStudent.Attendance}");

        Console.WriteLine($"Will Pass: {result.Passed}, Probability: {result.Probability:P2}");

    }

    }