tests/testthat/test_dcc.R

In treeclim: Numerical Calibration of Proxy-Climate Relationships

context("Numerical accuracy of dcc")

test_that("Numerical results are identical to benchmark version", {
  comp1 <- c(0.0138505849308184, -0.0322058662502632, -0.036688436243551, 
             0.049636664738515, 0.085956929428506, -0.0324165586695125, 0.14297109135422, 
             0.0898256720769509, 0.112778925299924, 0.0359701049303407, -0.130108125556049, 
             0.0702431963290987, 0.0423787305545529, 0.122101474377781, -0.00816049266647171, 
             0.0755044635317016, -0.0807748025288913, -0.022211825889583, 
             -0.112142811778103, -0.0284753243987742, 0.212301384996077, -0.108906724959197, 
             0.162539320482292, -0.0697467501498377, 0.133838281266717, 0.0172432788721393, 
             -0.153880361590093, 0.0267077261032693, -0.0277458214452862, 
             0.0444361269039689, -0.0856123210130883, 0.00357826949941429)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(42)
  coefs1 <- dcc(muc_spruce, muc_clim)$coef$coef
  expect_equal(coefs1, comp1)
  
  comp2 <- c(0.170430144507268, 0.117823349437864, -0.0537256677482294, 
             0.0661736844657929, 0.173529481497743, 0.0201817119502813, 0.0115497558775984)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(42)
  coefs2 <- dcc(rt_spruce, rt_prec, -4:-10)$coef$coef
  expect_equal(coefs2, comp2)
  
  comp3 <- c(-0.0392956369949715, 0.217670520229574, 0.184298376415361, 
             0.27766596206149, 0.259019290655206, 0.0842051584401493, 0.050908429942875, 
             -0.0249212641727796, -0.0657761954984126, -0.00751655658592378, 
             -0.0108474705692184, 0.0820489154839614)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(42)
  coefs3 <- dcc(rt_spruce, list(rt_prec, rt_temp), 4:9)$coef$coef
  expect_equal(coefs3, comp3)
  
  comp4 <- c(-0.0691539573105679, 0.295869404817051, 0.221174417023458, 
             0.369095273782355, 0.322312316185609, 0.124873739494498, 0.104793303567892, 
             -0.161605524443024, -0.174182562843724, -0.120592709112331, -0.14912796836124, 
             -0.0137164108335051)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(42)
  coefs4 <- dcc(rt_spruce, list(rt_prec, rt_temp), 4:9, method = "corr")$coef$coef
  expect_equal(coefs4, comp4)
  
  comp5 <- c(0.00616579407523024, 0.231072175537679, 0.245951917973043, 
             0.277349266793129, 0.343623537731027, 0.172419985883527)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(42)
  coefs5 <- dcc(rt_spruce, rt_prec, 4:9, method = "corr", moving = TRUE, win_size = 40)$coef$coef[,3]
  expect_equal(coefs5, comp5)
  
  comp6 <- c(0.155383432630196, 0.0918951748072474, -0.0638265717811419, 
             0.0710423074449692, 0.188587446576691, 0.0226450705463124, -0.0174483181998573
  )
  set.seed(42)
  coefs6 <- dcc(rt_spruce, rt_prec, -4:-10, boot = "exact")$coef$coef
  expect_equal(coefs6, comp6)
})