tests/testthat/test-uls_zones.R

In smerc: Statistical Methods for Regional Counts

context("test uls.zones matches PULSE example")

# Recreate examples from PULSE paper
# PULSE, progressive upper level set scan statistic for
# geospatial hotspot detection
# GP Patil, SW Joshi, RE Koli

# Table 1
cases <- c(90, 80, 80, 80, 80, 80, 70, 70, 60, 50, 40, 30)
pop <- rep(100, times = 12)
w <- matrix(0, nrow = 12, ncol = 12)
ubpop <- 1
w[1, c(2, 3, 8)] <- 1
w[2, c(1, 3, 7, 4)] <- 1
w[3, c(1, 2, 7, 8)] <- 1
w[4, c(2, 7)] <- 1
w[5, c(7, 6, 8)] <- 1
w[6, c(5, 12, 11)] <- 1
w[7, c(4, 2, 3, 5)] <- 1
w[8, c(1, 3, 5, 11)] <- 1
w[9, c(11, 10, 12)] <- 1
w[10, c(11, 9)] <- 1
w[11, c(8, 6, 9, 10)] <- 1
w[12, c(6, 9)] <- 1

z <- vector("list", 8)
z[[1]] <- 1
z[[2]] <- 1:4
z[[3]] <- 5:6
z[[4]] <- 1:8
z[[5]] <- 9
z[[6]] <- 9:10
z[[7]] <- 1:11
z[[8]] <- 1:12

uz <- uls.zones(cases, pop, w, ubpop = 1, check.unique = TRUE)

# example 9.1.1
cases2 <- c(50, 60, 40, 50, 30, 50)
pop2 <- rep(100, times = 6)
w2 <- matrix(0, nrow = 6, ncol = 6)
ubpop <- 1
w2[1, c(4, 2)] <- 1
w2[2, c(1, 3, 5)] <- 1
w2[3, c(2, 6)] <- 1
w2[4, c(1, 5)] <- 1
w2[5, c(4, 2, 6)] <- 1
w2[6, c(3, 5)] <- 1

z2 <- vector("list", 5)
z2[[1]] <- 2
z2[[2]] <- c(1, 2, 4)
z2[[3]] <- 6
z2[[4]] <- c(1:4, 6)
z2[[5]] <- 1:6

uz2 <- uls.zones(cases2, pop2, w2, ubpop = 1, check.unique = TRUE)
uz2 <- sapply(uz2, sort)

test_that("check accuracy of uls.zones", {
  expect_equal(z[1:4], uz[1:4])
  expect_equal(z2, uz2)
})