mathnet/mathnet-numerics

Math.NET Numerics is an open-source mathematics library for the .NET platform, written primarily in C#. It provides ready-to-use implementations of a wide range of numerical methods so developers do not have to write their own from scratch. The covered areas include linear algebra (matrix operations, decompositions), statistics, probability distributions, random number generation, interpolation, numerical integration, regression, curve fitting, and signal processing transforms like FFT. The library is part of the broader Math.NET initiative and targets .NET applications running on .NET 5 or later.NET Framework 4.6.1 or later, and .NET Standard 2.0. It installs via NuGet, the standard package manager for .NET projects, so adding it to a project is a single command or a click in an IDE. F# developers get specific support through a separate companion package that adds idiomatic extensions making the library feel more natural in functional code. A BigRational data type, originally from the F# PowerPack, is maintained here as well. For applications where raw performance matters, the default managed-code implementation behind linear algebra operations can be swapped out for bindings to optimized native libraries. Intel MKL and OpenBLAS are both supported through optional NuGet packages. This means the same code can run in pure managed mode during development or testing, and switch to native acceleration in production without any API changes. Optional data extension packages handle reading and writing matrices in text formats like CSV, the MatrixMarket format, and MATLAB's Level-5 .mat file format. These are separate installs, so you only add what your project needs. The library is released under the MIT license, which allows use in both open-source and commercial projects. Contributions are accepted, and the project has active discussion forums and a Stack Overflow presence.