Extreme Optimization > QuickStart Samples > Least Squares QuickStart Sample (C#)

Extreme Optimization QuickStart Samples

Least Squares QuickStart Sample (C#)

Illustrates solving least squares problems using the LeastSquaresSolver class (Extreme.Mathematics.Calculus namespace) in Visual Basic .NET.

VB.NET code Back to QuickStart Samples

using System;

namespace Extreme.Mathematics.QuickStart.CSharp
{
    // The GeneralMatrix and LeastSquaresSolver classes reside in the 
    // Extreme.Mathematics.LinearAlgebra namespace.
    using Extreme.Mathematics.LinearAlgebra;

    /// <summary>
    /// Illustrates solving least squares problems using the 
    /// LeastSquaresSolver class in the Extreme.Mathematics.LinearAlgebra 
    /// namespace of the Extreme Optimization Mathematics Library
    /// for .NET.
    /// </summary>
    class LinearEquations
    {
        /// <summary>
        /// The main entry point for the application.
        /// </summary>
        [STAThread]
        static void Main(string[] args)
        {
            // A least squares problem consists in finding
            // the solution to an overdetermined system of
            // simultaneous linear equations so that the
            // sum of the squares of the error is minimal.
            //
            // A common application is fitting data to a
            // curve. See the CurveFitting sample application
            // for a complete example.

            // Let's start with a general matrix. This will be
            // the matrix a in the left hand side ax=b:
            GeneralMatrix a = new GeneralMatrix(6, 4, new double[]
                {
                    1, 1, 1, 1, 1, 1,
                    1, 2, 3, 4, 5, 6,
                    1, 4, 9, 16, 25, 36,
                    1, 2, 1, 2, 1, 2
                });
            // Here is the right hand side:
            Vector b = new GeneralVector(1, 3, 6, 11, 15, 21);
            Matrix b2 = new GeneralMatrix(6, 2, new double[]
                {
                    1, 3, 6, 11, 15, 21,
                    1, 2, 3, 4, 5, 7
                });
            Console.WriteLine("a = {0}", a);
            Console.WriteLine("b = {0}", b);

            //
            // The LeastSquaresSolver class
            //

            // The following creates an instance of the
            // LeastSquaresSolver class for our problem:
            LeastSquaresSolver solver = new LeastSquaresSolver(a, b);
            // We can specify the solution method: normal
            // equations or QR decomposition. In most cases,
            // a QR decomposition is the most desirable:
            solver.SolutionMethod = LeastSquaresSolutionMethod.QRDecomposition;
            // The Solve method calculates the solution:
            Vector x = solver.Solve();
            Console.WriteLine("x = {0}", x.ToString("F4"));
            // The Solution property also returns the solution:
            Console.WriteLine("x = {0}", solver.Solution.ToString("F4"));
            // More detailed information is available from
            // additional methods.
            // The values of the right hand side predicted 
            // by the solution:
            Console.WriteLine("Predictions = {0}", solver.GetPredictions().ToString("F4"));
            // The residues (errors) of the solution:
            Console.WriteLine("Residuals = {0}", solver.GetResidues().ToString("F4"));
            // The total sum of squares of the residues:
            Console.WriteLine("Residual square error = {0}",
                solver.GetResidualSumOfSquares().ToString("F4"));

            //
            // Direct normal equations
            //

            // Alternatively, you can create a least squares
            // solution by providing the normal equations
            // directly. This may be useful when it is easy
            // to calculate the normal equations directly.
            // 
            // Here, we'll just calculate the normal equation:
            SymmetricMatrix aTa = SymmetricMatrix.FromOuterProduct(a);
            Vector aTb = b * a;
            // We find the solution by solving the normal equations
            // directly:
            x = aTa.Solve(aTb);
            Console.WriteLine("x = {0}", x.ToString("F4"));
            // However, properties of the least squares solution, such as
            // error estimates and residuals are not available.
            Console.Write("Press Enter key to exit...");
            Console.ReadLine();
        }
    }
}

Copyright 2004-2007, Extreme Optimization. All rights reserved.
Extreme Optimization, Complexity made simple, M#, and M Sharp  are trademarks of ExoAnalytics Inc.
Microsoft, Visual C#, Visual Basic, Visual Studio, and Visual Studio.NET are registered trademarks of Microsoft Corporation