Skip to contents

Numeric Exponential Kernel Specifications

exp_kernel_spec.Rd

Create a baseline-independent numeric kernel specification or extract one from an existing specification. These interfaces are experimental.

Usage

exp_kernel_spec(alpha, beta, eta = NULL, type_col_map = NULL)
as_exp_kernel(object)

Arguments

alpha, beta

Finite numeric matrices of equal positive dimensions, or scalars representing one by one matrices. Decay rates must be positive and equal within each row. Rows identify event types; columns identify kernels.

eta

Optional finite numeric matrix of the same shape.

type_col_map

One list entry per event type, containing distinct integer kernel columns. Entries can overlap. When omitted, the number of columns must equal the number of event types and event type i updates column i.

object

An hspec with numeric kernel parameters and no impact callback, or an existing exp_kernel_spec.

Details

Construction and conversion execute no user functions and use no random numbers. Function-valued kernel parameters and impact callbacks are rejected. Baselines, mark generators, mark densities and residual laws are not copied. Signed components and jumps are permitted; construction does not establish that the corresponding model can be simulated. Execution state is separate from the specification. Existing Hawkes calculation paths are unchanged.

Value

An S3 exp_kernel_spec containing alpha, beta, eta, dimens, kernel_cols and the resolved type_col_map.

See also

exp_kernel_replay

Examples

h <- new("hspec", mu = .3, alpha = .2, beta = 1)
kernel <- as_exp_kernel(h)
exp_kernel_replay(kernel, c(0, .25, .5), lambda_component0 = .1)
#> $kernel_integral
#>            [,1]
#> [1,] 0.00000000
#> [2,] 0.02211992
#> [3,] 0.10933729
#> 
#> $kernel_before
#>            [,1]
#> [1,] 0.10000000
#> [2,] 0.07788008
#> [3,] 0.16854279
#> 
#> $final_component
#>           [,1]
#> [1,] 0.3685428
#>