NMath Stats

NMath Stats is a premium, object-oriented statistical computing library developed by CenterSpace Software specifically for the .NET platform. It provides C#, F#, and VB.NET developers with production-grade data structures and functions. These tools allow developers to integrate high-performance math, engineering, scientific, and financial computations directly into their applications.

Note: Starting with NMath 7, CenterSpace unified all standalone statistical modules directly into the core CenterSpace.NMath.Core namespace. Key Capabilities and Features

The library expands the .NET ecosystem by replacing manual formula scripting with verified, mathematically stable algorithms across several domains:

Advanced Data Structures: Features a dedicated DataFrame class to handle mixed data types (numeric, string, boolean, datetime). It supports relational operations like appending, inserting, sorting, and permuting data.

Descriptive Statistics: Quickly extracts essential metrics including mean, variance, standard deviation, percentiles, quartiles, kurtosis, skewness, and root-mean-square (RMS).

Probability Distributions: Computes Probability Density Functions (PDF), Cumulative Distribution Functions (CDF), and inverse CDFs across dozens of distributions. Supported models include Normal (Gaussian), Weibull, Poisson, Chi-Square, and Student’s t.

Statistical Modeling & Regression: Full support for multiple linear regression, logistic regression (with odds ratios), Partial Least Squares (PLS), and Principal Component Analysis (PCA).

Hypothesis Testing: Out-of-the-box computation of p-values, critical values, and confidence intervals for standard parametric and non-parametric tests (z-test, t-test, F-test, Kolmogorov-Smirnov, and Kruskal-Wallis).

Analysis of Variance (ANOVA): Advanced tools for executing one-way and two-way ANOVA, balanced or unbalanced designs, and repeated measures (RANOVA). Performance: How It Accelerates Computation

NMath achieves native, low-latency execution speeds through deep hardware integration rather than relying purely on managed C# loops: 1. Native BLAS and LAPACK Bindings

The underlying math engine binds directly to vendor-optimized, machine-specific implementations of BLAS (Basic Linear Algebra Subprograms) and LAPACK (such as Intel MKL). This offloads heavy matrix math and linear algebra routine execution to raw C and Fortran speeds. 2. Data-View Design Pattern

To eliminate memory overhead, NMath uses a data-view design pattern. This allows multiple mathematical objects (like vectors and matrices) to point to the exact same block of contiguous memory without copying data. It greatly accelerates operations on massive real-time datasets. .NET Statistical Computation with NMath Stats