Package dev.nm

NMDEV numerical computing: a guide to the public API.

See: Description

Package dev.nm Description

NMDEV numerical computing: a guide to the public API.

Start with a task

• Construct matrices and multiply vectors: DenseMatrix.
• Understand vector constructors and storage: DenseVector.
• Summarize samples: Mean and Variance.
• Compute probabilities and quantiles: NormalDistribution.
• Find a scalar root: BrentRoot.
• Minimize an objective: NelderMeadMinimizer.
• Solve a quadratic program: QPSimpleMinimizer.
• Solve linear least squares: OLSSolver.
• Interpolate observations: LinearInterpolation.
• Integrate a function: Simpson.
• Reproduce a random stream: UniformRNG.
• Fit and predict a regression: OLSRegression.
• Integrate an initial-value ODE: EulerMethod.
• Model extreme values: GeneralizedEVD.
• Calculate price returns: ReturnsCalculators.

Before running an example

Add the combined NMDEV dependency to your application's classpath and provide a valid NMDEV license covering the packages used. For the standard file-based setup, put your supplied nm.lic in the JVM working directory. The example statements go inside a method such as public static void main(String[] args) throws Exception; import statements go at the top of the source file. The examples use Java assertions to show expected results; run with java -ea to check them.

Conventions shared by the examples

Matrix and vector entry indices start at one. Java arrays start at zero. Check constructor ownership rules: some inputs are retained rather than copied. Read-only wrappers do not by themselves make snapshots. Match dimensions, units, tolerances, and probability domains to the operation being called. Numerical tolerances are algorithm-specific; verify residuals and convergence.