Crate nabled_ml 

Ndarray-native ML-oriented numerical domains for the nabled workspace.

nabled-ml composes nabled-linalg kernels into higher-level ML/statistical routines over ndarray vectors and matrices.

Included Domains

1. iterative: iterative linear-system solvers.
2. optimization: first/second-order optimization routines.
3. jacobian: numerical Jacobian/gradient/Hessian estimators.
4. pca: principal component analysis and transforms.
5. regression: linear regression routines.
6. stats: covariance/correlation/centering utilities.

Feature Flags

1. blas: enables BLAS acceleration via nabled-linalg/blas.
2. openblas-system: enables provider-backed LAPACK paths via system OpenBLAS.
3. openblas-static: enables provider-backed LAPACK paths via statically linked OpenBLAS.
4. netlib-system: enables provider-backed LAPACK paths via system Netlib LAPACK.
5. netlib-static: enables provider-backed LAPACK paths via statically linked Netlib LAPACK.

Example

use ndarray::{arr1, arr2};
use nabled_ml::regression;

let x = arr2(&[[1.0_f64, 1.0], [1.0, 2.0], [1.0, 3.0]]);
let y = arr1(&[1.0_f64, 2.0, 3.0]);
let model = regression::linear_regression(&x, &y)?;
assert_eq!(model.coefficients.len(), 2);

Modules

iterative

Iterative linear system solvers over ndarray matrices.

jacobian

Numerical Jacobian/gradient/Hessian computation over ndarray vectors.

optimization

Optimization primitives over ndarray vectors.

pca

Principal component analysis over ndarray matrices.

regression

Linear regression over ndarray matrices.

stats

Statistical utilities over ndarray matrices.