crps_vonmises: Circular CRPS

In circtree: Regression Trees and Forests for Circular Responses

Description

Continuous Ranked Probability Score (CRPS) for a circular response following the von Mises distribution.

Usage

  crps_vonmises(y, mu, kappa, sum = FALSE)

Arguments

y	numeric. Circular response.
mu	numeric. Location parameter of the von Mises distribution.
kappa	numeric. Concentration parameter of the von Mises distribution.
sum	logical. Should the sum of the CRPS-values over all response values be returned.

Value

For sum=TRUE the sum of the CRPS-values of all response values is returned. Otherwise a vector of the same length as y with the observation-wise CRPS-values is returned.

See Also

circtree

Examples

  set.seed(123)
  sdat <- circtree_simulate(n = 100)
  m1.circtree <- circtree(y ~ x1 + x2, data = sdat)
  foo <- function(x, deg = FALSE){
    if(deg) x <- x *  pi / 180
    tmp <- x 
    tmp <- ifelse(tmp > pi, -(pi - (tmp - pi)), tmp)
    if(deg) tmp <- tmp *  180 / pi
    tmp
  }
   
  testcrps <- crps_vonmises(foo(sdat$y), mu = predict(m1.circtree)$mu, predict(m1.circtree)$kappa) 

circtree documentation built on Aug. 14, 2019, 3 p.m.