Regression 1.1.0
See the version list below for details.
dotnet add package Regression --version 1.1.0
NuGet\Install-Package Regression -Version 1.1.0
<PackageReference Include="Regression" Version="1.1.0" />
<PackageVersion Include="Regression" Version="1.1.0" />
<PackageReference Include="Regression" />
paket add Regression --version 1.1.0
#r "nuget: Regression, 1.1.0"
#:package Regression@1.1.0
#addin nuget:?package=Regression&version=1.1.0
#tool nuget:?package=Regression&version=1.1.0
Библиотека работы с статистикой
AlphaPascal.Math.Regression
Polynomial regression/approximation and interpolation.
Two-factor polynomial fit (X1, X2 → Y)
Both solvers implement IPolynomialFitService (GetValues(IEnumerable<DataTwoFact>)), take an
IBasisExponents (SecondOrderBasisExponents / ThirdOrderBasisExponents /
FourthOrderBasisExponents, 9 / 16 / 25 coefficients) and return the coefficients of the
original monomials X1^i · X2^j in that basis's fixed order, so they are interchangeable:
Func<List<DataTwoFact>, IEnumerable<double>> leastSquares =
data => new PolynomialLeastSquaresSolver(new ThirdOrderBasisExponents()).GetValues(data);
Func<List<DataTwoFact>, IEnumerable<double>> minimax =
data => new MinimaxPolynomialSolver(new ThirdOrderBasisExponents()).GetValues(data);
| Solver | Minimizes | Use when |
|---|---|---|
PolynomialLeastSquaresSolver (also ILeastSquaresRegressionService) |
sum of squared errors | Default. Accuracy matters between calibration temperatures, or the data may contain noisy points. |
MinimaxPolynomialSolver |
maximum absolute error over the given points | Acceptance is a max-error bound checked strictly at the calibration points. |
MinimaxPolynomialSolver is iterative (Lawson's reweighted least squares, maxIterations default
1000) and lands within a few percent of the exact minimax optimum. It lowers the in-sample max error
(third order on the 223/224 datasets: 0.00152 → 0.00095 and 0.00144 → 0.00104), but it chases the
worst points: a noisy point pulls the whole fit toward itself, and it predicts unseen temperatures
worse than least squares (leave-one-series-out, worst interior series: 0.0019 → 0.0029 and
0.0032 → 0.0048).
For either solver, the number of distinct temperatures must exceed the temperature degree of the basis: at least 3 / 4 / 5 for second / third / fourth order. Otherwise the high-power terms are undetermined, and the in-sample error will not show it.
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net8.0 is compatible. net8.0-android was computed. net8.0-browser was computed. net8.0-ios was computed. net8.0-maccatalyst was computed. net8.0-macos was computed. net8.0-tvos was computed. net8.0-windows was computed. net9.0 was computed. net9.0-android was computed. net9.0-browser was computed. net9.0-ios was computed. net9.0-maccatalyst was computed. net9.0-macos was computed. net9.0-tvos was computed. net9.0-windows was computed. net10.0 was computed. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net8.0
- MathNet.Symbolics (>= 0.25.0)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.