random_labelling_test: Random labeling test
In myllym/spptest: Spatial point process testing

Description Usage Arguments Details Examples

Do a random labelling deviation test for a pattern using K_f- or L_f-functions.

random_labelling_test(pattern, mtf_name = "m", nsim = 999L, r_min = NULL,
  r_max = NULL, r_vec = NULL, measure = "max", scaling = "qdir",
  use_L = TRUE, edge_correction = "translate", use_theo = TRUE,
  method = "permute", use_biased_lambda2 = FALSE, ties = "random")

`pattern`	A `ppp` object as the simple marked point pattern to be analysed. The marks need to be in the form of a numeric vector. The window has to have the type "rectangle".
`mtf_name`	A vector of mark test function names. "1" stands for the unmarked K-function. Accepted values are '1', 'm', 'mm', 'gamma', 'gammaAbs' and 'morAbs'. See `summ_func_random_labelling`.
`nsim`	The number of permutations.
`r_min`	The minimum radius to include.
`r_max`	A positive scalar value representing the maximum radius that should be considered. r_vec overrides r_max in calculation of functions. However, after calculation of functions, they are cropped to the interval [r_min, r_max] (if max(r_vec) > r_max loss in computing time). By default, r_max is NULL and will get a sensible default.
`r_vec`	A monotonically increasing vector of non-negative r-values to act as the endpoints of the bins for the K_f-functions. The bins are exclusive on the left and inclusive on the right. If the first vector element has value zero, it will be regarded as the collapsed bin [0, 0], and the next bin will start from and exclude 0.
`measure`	The deviation measure to use. Default is 'max'. Must be one of the following: 'max', 'int', 'int2'.
`scaling`	The name of the scaling to use. Options include 'none', 'q', 'qdir' and 'st'. 'qdir' is default.
`use_L`	A boolean describing whether the L_f function, L_f = sqrt(K_f/pi), should be used instead of the K_f function.
`edge_correction`	The name of the edge correction to be used. Options are 'translate' and 'none'.
`use_theo`	Whether to use the theoretical summary function or the mean of the simulations.
`method`	The name of the method to create simulations under the null hypothesis. 'permute' results in permutations of the marks. Using 'sample' will sample the marks from the empirical mark distribution with replacement. 'permute' is the default.
`use_biased_lambda2`	A logical scalar on whether to use the biased or the unbiased (in the Poisson case) estimate of the intensity squared.
`ties`	The method to treat tied values with. If one or more of the elements of sim_vec are equal to obs, how should the rank of obs be determined? For 'conservative' the resulting p-value will be the highest possible. For 'liberal' the p-value will be the lowest possible. For 'random' the rank of the obs within the tied values is uniformly sampled so that the resulting p-value is at most the conservative option and at least the liberal option. For 'midrank' the mid-rank within the tied values is taken. 'midrank' is the default.

Given a marked point pattern, the function performs simulations under the random labelling hypothesis, estimates K_f- or L_f-functions for the pattern and simulations and makes a deviation test.

require(spatstat)
mpp <- spruces
# T(r) = \hat{L}_m(r), an estimator of the L_m(r) function
p <- random_labelling_test(mpp, mtf_name = 'm', nsim=999, r_min=0, r_max=9.5)
p