Math.NET Numerics Features Overview

You can suggest or vote for new features or ideas here.

Core¶

Namespace: MathNet.Numerics

Complex Numbers (including trigonometric functions)
Mathematical & Scientific Constants and Prefixes (2007 CODATA)
Precision class, for safe floating point handling
Various helper and data structure classes

Trigonometry Functions¶

Namespace: MathNet.Numerics
Class: MathNet.Numerics.Trig

Real and complex trigonometry functions

Conversion between degree, radian and grad.
Sine, Cosine, Tangent, Cotangent, Secant, Cosecant
Inverse: Sine, Cosine, Tangent, Cotangent, Secant, Cosecant
Hyperbolic: Sine, Cosine, Tangent, Cotangent, Secant, Cosecant
Inverse Hyperbolic: Sine, Cosine, Tangent, Cotangent, Secant, Cosecant

Number Theory¶

Namespace: MathNet.Numerics.NumberTheory
Class: MathNet.Numerics.NumberTheory.IntegerTheory

IsEven, IsOdd
IsPowerOfTwo
PowerOfTwo, CeilingToPowerOfTwo
IsPerfectSquare
GreatestCommonDivisor (gcd) of two or more integers
ExtendedGreatestCommonDivisor (egcd) of two integers
LeastCommonMultiple (lcm) of two or more integers

Numerically Stable Functions¶

Namespace: MathNet.Numerics
Class: MathNet.Numerics.SpecialFunctions

Hypotenuse: (a,b) -> sqrt(a^2+b^2)
ExponentialMinusOne: x -> exp(x)-1

Special Functions¶

Namespace: MathNet.Numerics
Class: MathNet.Numerics.Fn

Logarithmic Factorial
Factorial
Logarithmic Binomial Coefficient
Binomial Coefficient

Logarithmic Gamma
Gamma (supports negative numbers as well)
Regularized Gamma

Digamma (Psi)
Inverse Digamma

Logarithmic Beta
Regularized Beta

Error Function (erf) and complement (erfc)
Inverse Error Function and complement

Combinatorics ¶

Namespace: MathNet.Numerics
Class: MathNet.Numerics.Combinatorics

Counting: Variations, Variations with repetition, Combinations, Combinations with repetition, Permutations

Probability Distributions¶

Namespace: MathNet.Numerics.Distributions

Continuous Probability Distributions¶

Continuous probability distributions support both the probability density function (pdf) and the cumulative distribution function (cdf), as well as the usual probability parameters. Additionally, random numbers can be generated based on the configured probability model parameters and some random number source.

Uniform
Normal (Gaussian with mean and variance)
Gamma
Beta
Weibull

Discrete Probability Distributions¶

Discrete probability distributions support both the probability mass function (pmf) and the cumulative distribution function (cdf), as well as the usual probability parameters. Additionally, random numbers can be generated based on the configured probability model parameters and some random number source.

Bernoulli

Multivariate Probability Distributions¶

Multinomial
Dirichlet

Random Sources¶

Namespace: MathNet.Numerics.Random

All implementations inherit the .Net framework provided System.Random class for interoperability.

Mersenne Twister

Note that random sources should be reused, so be careful to create only one instance (per thread) and share it internally.

Interpolation¶

Namespace: MathNet.Numerics.Interpolation

Most interpolation algorithms also support numeric differentiation and integration. A facade class Interpolate is provided for easy access, but if needed the algorithms can also be used directly in the Algorithms sub-namespace. All implementations implement the interface IInterpolation.

Rational pole-free, on arbitrary points (Barycentric Floater-Hormann Algorithm)
Rational with poles, on arbitrary points (Bulirsch & Stoer Algorithm)
Neville Polynomial, on arbitrary points (Neville Algorithm)
Polynomial, on equidistant points (Barycentric Algorithm)
Linear Spline, on arbitrary points
Cubic Spline, with boundary conditions on arbitrary points
Natural Cubic Spline, on arbitrary points
Akima Cubic Spline, on arbitrary points
Custom Barycentric Interpolation, based on provided barycentric weights
Custom Spline Interpolation, based on provided spline coefficients
Custom Cubic Hermite Spline Interpolation, based on provided derivatives

If unsure what to choose, we recommend to simply use Interpolate.Common(x,y) which internally uses the barycentric rational pole-free interpolation.

Code Sample¶

double[] t = new double[] { -2.0, -1.0, 0.0, 1.0, 2.0 };
double[] x = new double[] { 1.0, 2.0, -1.0, 0.0, 1.0 };
IInterpolation interp = Interpolate.RationalWithoutPoles(t, x);
double a = interp.Interpolate(-0.5);

Linear Algebra¶

Namespace: MathNet.Numerics.LinearAlgebra

Vector: Dense Real Double

Integral Transforms¶

Namespace: MathNet.Numerics.IntegralTransforms

Complex Fast Fourier Transformation
Real Hartley Transform

The transformation behavior can be configured (scaling, exponent sign, etc). See here for more details and code samples around Fourier transforms.