Struct ArdRbfKernel 

pub struct ArdRbfKernel { /* private fields */ }

ARD (Automatic Relevance Determination) RBF kernel.

K(x, y) = σ² * exp(-0.5 * Σ((x_i - y_i)² / l_i²))

Each dimension has its own length scale l_i, allowing the kernel to automatically weight features by their relevance.

Implementations

impl ArdRbfKernel

pub fn new(length_scales: Vec<f64>) -> Result<Self>

Create a new ARD RBF kernel with per-dimension length scales.

Arguments

• length_scales - Length scale for each dimension (all must be positive)

Example

use tensorlogic_sklears_kernels::ard_kernel::ArdRbfKernel;

let kernel = ArdRbfKernel::new(vec![1.0, 2.0, 0.5]).unwrap();

pub fn with_variance(length_scales: Vec<f64>, variance: f64) -> Result<Self>

Create ARD RBF kernel with custom signal variance.

Arguments

• length_scales - Per-dimension length scales
• variance - Signal variance (output scale, must be positive)

pub fn length_scales(&self) -> &[f64]

Get the length scales.

pub fn variance(&self) -> f64

Get the signal variance.

pub fn ndim(&self) -> usize

Get the number of dimensions.

pub fn compute_gradient(&self, x: &[f64], y: &[f64]) -> Result<KernelGradient>

Compute the kernel gradient with respect to hyperparameters.

Returns gradients for:

1. Each length scale (one per dimension)
2. The signal variance

This is useful for hyperparameter optimization via gradient descent.

Trait Implementations

impl Clone for ArdRbfKernel

fn clone(&self) -> ArdRbfKernel

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ArdRbfKernel

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Kernel for ArdRbfKernel

fn compute(&self, x: &[f64], y: &[f64]) -> Result<f64>

Compute kernel value between two inputs.

fn name(&self) -> &str

Get kernel name for identification.

fn compute_matrix(&self, inputs: &[Vec<f64>]) -> Result<Vec<Vec<f64>>>

Compute kernel matrix for a set of inputs.

fn is_psd(&self) -> bool

Check if kernel is positive semi-definite.