Numerical Computing in .NET with NumSharp

NumSharp is an open-source library for .NET designed to facilitate numerical computing and data analysis, emulating the functionality of NumPy, one of the most popular libraries in the Python ecosystem.

NumSharp provides a simple API for manipulating arrays, performing mathematical calculations, and working with numerical data in .NET applications.

Features of NumSharp:

• N-dimensional Arrays: Support for multidimensional arrays, similar to NumPy arrays.
• Mathematical Operations: Functions for linear algebra, statistics, and general mathematical operations.
• Interoperability with ML.NET: Facilitates integration with ML.NET for machine learning model development.
• Compatibility: Works in .NET Core and .NET Framework projects.

Installing NumSharp

You can install NumSharp via the NuGet package manager. Open your terminal or the NuGet Package Manager console in Visual Studio and run the following command:

Install-package NumSharp

Or through the NuGet interface in Visual Studio, search for NumSharp and install it in your project.

How to Use NumSharp

Creating Arrays and Matrices

NumSharp provides an NDArray class for working with multidimensional arrays. Here is a basic example of how to create and manipulate arrays:

using NumSharp;

class Program
{
    static void Main(string[] args)
    { 
		// Create a one-dimensional array
		var array1D = np.array(new int[] { 1, 2, 3, 4, 5 });

		// Create a two-dimensional array
		var array2D = np.array(new int[,] { { 1, 2 }, { 3, 4 } });

		// Display the arrays
		Console.WriteLine("Array 1D:");
		Console.WriteLine(array1D.ToString());

		Console.WriteLine("Array 2D:");
		Console.WriteLine(array2D.ToString());
    }
}

Array 1D:
[1, 2, 3, 4, 5]
Array 2D:
[[1, 2],
[3, 4]]

Basic Mathematical Operations

NumSharp includes a series of useful mathematical functions for basic operations on arrays and matrices:

using NumSharp;

class Program
{
    static void Main(string[] args)
    {
        var array = np.array(new double[] { 1, 2, 3, 4, 5 });

        // Mathematical operations
        var squared = np.power(array, 2); // Raises each element to the power of two
        var sum = np.sum(array); // Sums all elements

        Console.WriteLine("Original Array:");
        Console.WriteLine(array.ToString());

        Console.WriteLine("Squared Array:");
        Console.WriteLine(squared.ToString());

        Console.WriteLine("Sum of Elements:");
        Console.WriteLine(sum.ToString());
    }
}

Data Manipulation

NumSharp allows for complex operations on matrices, such as transposition, reshaping, and advanced mathematical operations:

using NumSharp;

class Program
{
    static void Main(string[] args)
    {
        // Create a 3x3 matrix
        var matrix = np.array(new double[,] {
            { 1, 2, 3 },
            { 4, 5, 6 },
            { 7, 8, 9 }
        });

        // Transpose the matrix
        var transpose = np.transpose(matrix);

        // Reshape the matrix to 1x9
        var reshape = np.reshape(matrix, new Shape(1, 9));

        Console.WriteLine("Original Matrix:");
        Console.WriteLine(matrix.ToString());

        Console.WriteLine("Transposed Matrix:");
        Console.WriteLine(transpose.ToString());

        Console.WriteLine("Reshaped Matrix:");
        Console.WriteLine(reshape.ToString());
    }
}

Integration with ML.NET

NumSharp is compatible with ML.NET, making it easier to work with numerical data in the context of machine learning:

using NumSharp;
using System;
using System.Linq;

class Program
{
    static void Main(string[] args)
    {
        // Create a feature array
        var features = np.array(new double[] { 1.0, 2.0, 3.0, 4.0, 5.0 });

        // Use NumSharp to prepare data for ML.NET
        var normalizedFeatures = (features - np.min(features)) / (np.max(features) - np.min(features));

        Console.WriteLine("Normalized Features:");
        Console.WriteLine(normalizedFeatures.ToString());
    }
}

For more information and to explore the complete documentation, visit the official NumSharp repository on GitHub. Here you will find additional examples, tutorials, and resources.