Struct IndexKernel 

pub struct IndexKernel { /* private fields */ }

Index kernel: K(i, j) = B[i, j] where B is task covariance matrix.

Pure task-based similarity without feature component. Useful as a building block for more complex multi-task kernels.

Implementations

impl IndexKernel

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

Create an index kernel from task covariance matrix.

The covariance matrix should be symmetric positive semi-definite.

pub fn identity(num_tasks: usize) -> Result<Self>

Create with identity covariance (independent tasks).

pub fn uniform(num_tasks: usize, correlation: f64) -> Result<Self>

Create with uniform covariance (all tasks equally similar).

pub fn get_task_covariance(&self, task_i: usize, task_j: usize) -> Result<f64>

Get task covariance value.

pub fn num_tasks(&self) -> usize

Get number of tasks.

pub fn covariance_matrix(&self) -> &Vec<Vec<f64>>

Get the full covariance matrix.

Trait Implementations

impl Clone for IndexKernel

fn clone(&self) -> IndexKernel

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for IndexKernel

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

Formats the value using the given formatter.