Releases · dcwuser/metanumerics
- Int128 and UInt128 extended precision integer types
- DoubleDouble supports infinity and NaN
- LogGamma to DoubleDouble accuracy
- Complete elliptic integral of the 3rd kind
- Bessel and Airy zeros, scaled modified Bessel functions, hyperbolic Shi and Cin functions
- Improved speed, accuracy, and handling of extreme values for many advanced functions
- Noncentral chi squared distribution
- Faster inverse CDF and random deviate generation for many discrete distributions
- Added abstract RegressionResult class to cover all results that return Residuals
- Targets .NET Standard 1.1 (.NET Framework 4.5 or higher or .NET Core)
- Data frame system for data import, manipulation, and analysis.
- Shapiro-Francia test for normality.
- Hypergeometric, Frechet, and other additional distributions.
- Gauss hypergeometric 2F1 function for real arguments.
- Double-double type for extended precision floating point computations.
- Specific types for fit results, analysis settings and results. Static extension methods for statistical analysis. Improved matrix decomposition API.