KernelComposed {BKTR}

R6 class for Composed Kernels

Description

R6 class for Composed Kernels

Super class

BKTR::Kernel -> KernelComposed

Public fields

name

The kernel's name

parameters

The parameters of the kernel (list of KernelParameter)

left_kernel

The left kernel to use for composition

right_kernel

The right kernel to use for composition

composition_operation

The operation to use for composition

has_dist_matrix

Identify if the kernel has a distance matrix or not

Methods

Public methods

• KernelComposed$new()

• KernelComposed$core_kernel_fn()

• KernelComposed$set_positions()

• KernelComposed$clone()

Inherited methods

• BKTR::Kernel$add_jitter_to_kernel()
• BKTR::Kernel$kernel_gen()
• BKTR::Kernel$plot()

Method new()

Create a new KernelComposed object.

Usage

KernelComposed$new(left_kernel, right_kernel, new_name, composition_operation)

Arguments

left_kernel

Kernel: The left kernel to use for composition

right_kernel

Kernel: The right kernel to use for composition

new_name

String: The name of the composed kernel

composition_operation

CompositionOps: The operation to use for composition

Method core_kernel_fn()

Method to compute the core kernel's covariance matrix

Usage

KernelComposed$core_kernel_fn()

Returns

The core kernel's covariance matrix

Method set_positions()

Method to set the kernel's positions and compute the distance matrix

Usage

KernelComposed$set_positions(positions_df)

Arguments

positions_df

Dataframe: The positions of the points in a dataframe format

Returns

NULL, set the kernel's positions and compute the distance matrix

Method clone()

The objects of this class are cloneable with this method.

Usage

KernelComposed$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Create a new locally periodic kernel
k_loc_per <- KernelComposed$new(
  left_kernel = KernelSE$new(),
  right_kernel = KernelPeriodic$new(),
  new_name = 'Locally Periodic Kernel',
  composition_operation = CompositionOps$MUL
)
# Set the kernel's positions
positions_df <- data.frame(x=c(-4, 0, 3), y=c(-2, 0, 2))
k_loc_per$set_positions(positions_df)
# Generate the kernel's covariance matrix
k_loc_per$kernel_gen()