Struct LMCKernel 

pub struct LMCKernel { /* private fields */ }

Linear Model of Coregionalization (LMC) kernel.

K((x, i), (y, j)) = Σ_q B_q[i, j] * k_q(x, y)

where:

• Each (B_q, k_q) pair represents a latent process
• B_q is a task covariance matrix
• k_q is a kernel function

LMC is more expressive than ICM as it allows different kernels to capture different aspects of task relationships.

Implementations

impl LMCKernel

pub fn new(num_tasks: usize) -> Self

Create a new LMC kernel.

pub fn add_component( &mut self, kernel: Box<dyn Kernel>, task_covariance: Vec<Vec<f64>>, ) -> Result<()>

Add a latent process component.

pub fn compute_tasks(&self, x: &TaskInput, y: &TaskInput) -> Result<f64>

Compute LMC kernel value for task inputs.

pub fn num_components(&self) -> usize

Get number of components.

pub fn num_tasks(&self) -> usize

Get number of tasks.

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

Compute full kernel matrix for multiple task inputs.