Struct GaussianProcessModel 

pub struct GaussianProcessModel {
    pub x_train: Vec<Vec<f64>>,
    pub y_train: Vec<f64>,
    pub config: GaussianProcessConfig,
    pub hyperparameters: GPHyperparameters,
    /* private fields */
}

Gaussian Process regression model.

Conditioned on training data, this model provides posterior mean and variance predictions via an exact Cholesky-based solve of K + σ²I.

Fields

x_train: Vec<Vec<f64>>

Training input data

y_train: Vec<f64>

Training output data

config: GaussianProcessConfig

GP configuration

hyperparameters: GPHyperparameters

Learned hyperparameters

Implementations

impl GaussianProcessModel

pub fn new( x_train: Vec<Vec<f64>>, y_train: Vec<f64>, config: GaussianProcessConfig, ) -> BayesianOptResult<Self>

Create new Gaussian Process model

pub fn fit(&mut self) -> BayesianOptResult<()>

Fit the Gaussian Process model.

Sets heuristic hyperparameters, then Cholesky-factorizes K + σ²I and precomputes α for fast predictive-mean evaluation.

pub fn predict(&self, x: &[f64]) -> BayesianOptResult<(f64, f64)>

Predict the posterior mean and variance at a new point.

Implements the exact-GP predictive equations: μ(x*) = m(x*) + k*ᵀ α and σ²(x*) = k(x*, x*) − vᵀ v with v = L⁻¹ k*.

pub fn log_marginal_likelihood(&self) -> BayesianOptResult<f64>

Exact log marginal likelihood of the training data under the fitted GP.

log p(y) = −½ (y − m)ᵀ α − Σ ln L_ii − ½ n ln(2π), where the middle term equals ½ ln|K + σ²I| because L is the Cholesky factor.

Trait Implementations

impl Clone for GaussianProcessModel

fn clone(&self) -> GaussianProcessModel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for GaussianProcessModel

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

Formats the value using the given formatter.