Local G*

Calculate the local Gi* statistic.

Usage

local_gstar(x, nb, wt, alternative = "two.sided", ...)

local_gstar_perm(x, nb, wt, nsim = 499, alternative = "two.sided", ...)

Arguments

x

A numeric vector.

nb

a neighbor list object for example as created by st_contiguity().

wt

a weights list as created by st_weights().

alternative

default "two.sided". Should be one of "greater", "less", or "two.sided" to specify the alternative hypothesis.

...

methods passed to spdep::localG() or spdep::localG_perm()

nsim

The number of simulations to run.

Value

a data.frame with columns:

• gi: the observed statistic

• e_gi: the permutation sample mean

• var_gi: the permutation sample variance

• p_value: the p-value using sample mean and standard deviation

• p_folded_sim: p-value based on the implementation of Pysal which always assumes a two-sided test taking the minimum possible p-value

• skewness: sample skewness

• kurtosis: sample kurtosis

Examples

nb <- st_contiguity(guerry)
wt <- st_weights(nb)
x <- guerry$crime_pers

res <- local_gstar_perm(x, nb, wt)
head(res)
#>     gi_star cluster       e_gi       var_gi    std_dev    p_value p_sim
#> 1  1.342359    High 0.01281212 2.828987e-06  0.8669055 0.38599380 0.408
#> 2  2.595578    High 0.01231754 2.534491e-06  2.1922882 0.02835870 0.036
#> 3  2.388872    High 0.01226355 2.355173e-06  2.0995950 0.03576448 0.044
#> 4 -1.848187     Low 0.01098208 2.788616e-06 -1.5970973 0.11024403 0.120
#> 5 -1.209736     Low 0.01146742 3.231416e-06 -1.2477284 0.21213055 0.212
#> 6 -2.034142     Low 0.01097455 1.880792e-06 -1.5710781 0.11616450 0.120
#>   p_folded_sim    skewness    kurtosis
#> 1        0.204  0.11601646 -0.37505144
#> 2        0.018  0.13538275 -0.07345225
#> 3        0.022  0.07291573 -0.24293931
#> 4        0.060 -0.05713113 -0.15455315
#> 5        0.106  0.13602702 -0.40875762
#> 6        0.060  0.05456544 -0.27237752

res <- local_gstar(x, nb, wt)
head(res)
#> [1]  1.342359  2.595578  2.388872 -1.848187 -1.209736 -2.034142