# test replicate stratification option context("Replicate") test_that("Replicates: Savannah sparrows", { skip_on_cran() data("Savannah_sparrow_1981") ss81 <- Savannah_sparrow_1981 # relevel data for distance comparability ss81$Region.Label <- relevel(ss81$Region.Label, ref=4) cf <- convert_units("meter", NULL, "hectare") #it <- ds(ss81, key="hr", formula = ~Region.Label, convert.units = cf, # transect="point", scale ="width", # initial=list(scale=c(log(32.18), -0.1456, 0.2502, 0.8775E-01), # shape=log(4.549))) ss81d <- unflatten(ss81)$data # fit in ddf to fix initial values result <- ddf(dsmodel=~mcds(key="hr", formula=~1), data=ss81d, method="ds", meta.data=list(width=max(ss81d$distance), point=TRUE), control=list(initial=list(scale=log(33.23), shape=log(4.262)), nofit=TRUE)) # for density estimation ss81$Area <- 0 what <- dht2(result, flatfile = ss81, convert_units = cf, strat_formula=~Region.Label, stratification = "replicate") expect_equal(what$Abundance, c(0.63466, 0.67093, 1.1424, 0.92479, 0.84319), tol=1e-3) expect_equal(what$Abundance_CV, c(18.88, 19.78, 16.67, 17.09, 16.88)/100, tol=1e-3) expect_equal(what$df[1:4], c(163.41, 156.26, 187.75, 182.20), tol=1e-1) expect_equal(what$df[5], 6.16, tol=1e-2) expect_equal(what$LCI, c(0.43861, 0.45567, 0.82406, 0.66168, 0.56098), tol=1e-3) expect_equal(what$UCI, c(0.91835, 0.98788, 1.5837, 1.2925, 1.2674), tol=1e-3) # Distance results # Estimate %CV df 95% Confidence Interval # ------------------------------------------------------ # Stratum: PASTURE 0 # Hazard/Cosine # D 0.63466 18.88 163.41 0.43861 0.91835 # N 0.00000 # Stratum: PASTURE 1 # Hazard/Cosine # D 0.67093 19.78 156.26 0.45567 0.98788 # N 0.00000 # Stratum: PASTURE 2 # Hazard/Cosine # D 1.1424 16.67 187.75 0.82406 1.5837 # N 0.00000 # Stratum: PASTURE 3 # Hazard/Cosine # D 0.92479 17.09 182.20 0.66168 1.2925 # N 0.00000 # # Pooled Estimates: # Estimate %CV df 95% Confidence Interval # ------------------------------------------------------ # D 0.84319 16.88 6.16 0.56098 1.2674 # N 0.00000 # })