tests/testthat/test-moparam.R

In hsloot/rmo: A package for the Marshall-Olkin distribution

d <- 10L

test_that("ex_intensities parameter is calculated correctly", {
  bf <- AlphaStableBernsteinFunction(alpha = 0.1033993)

  expect_equal(
    exIntensities(bf, d),
    sapply(1:d, function(i) valueOf(bf, d - i, i, d, i))
  )

  uex_intensities <- uexIntensities(bf, d)
  expect_equal(
    exIntensities(bf, d),
    sapply(
      1:d,
      function(i) {
        multiply_binomial_coefficient(uex_intensities[[i]], d, i)
      }
    )
  )
})

test_that("intensities parameter is calculated correctly", {
  bf <- AlphaStableBernsteinFunction(alpha = 0.07608632)

  tmp <- sapply(1:d, function(i) valueOf(bf, d - i, i))
  intensities <- numeric(2^d - 1)
  for (j in seq_along(intensities)) {
    count <- 0
    for (i in 1:d) {
      count <- count + Rcpp__is_within(i, j)
    }
    intensities[j] <- tmp[count]
  }
  expect_equal(intensities(bf, d), intensities)
})

test_that("ex_qmatrix parameter is calculated correctly", {
  bf <- AlphaStableBernsteinFunction(alpha = 0.6101982)

  ex_qmatrix <- matrix(0, nrow = d + 1, ncol = d + 1)
  ex_qmatrix[1, -1] <- exIntensities(bf, d)
  for (i in 1:d) {
    if (i < d) {
      for (j in (i + 1):d) {
        ex_qmatrix[1 + i, 1 + j] <-
          (d - j + 1) / (d - i + 1) * ex_qmatrix[i, j] +
          (j + 1 - i) / (d - i + 1) * ex_qmatrix[i, j + 1]
      }
    }
  }
  diag(ex_qmatrix) <- -apply(ex_qmatrix, 1, sum)
  expect_equal(exQMatrix(bf, d), ex_qmatrix)

  expect_equal(exQMatrix(bf, d), exQMatrix(bf, d + 1)[-1, -1])
})

test_that("sum(ex_intensities) is calculated correctly (base case)", {
  bf <- AlphaStableBernsteinFunction(alpha = 0.8598596)

  expect_equal(
    sum(exIntensities(bf, d)),
    valueOf0(bf, d)
  )
})

test_that("sum(ex_intensities) is calculated correctly (corner case)", {
  bf <- AlphaStableBernsteinFunction(log2(2 - 99.999e-2))
  d <- 125

  expect_equal(
    sum(exIntensities(bf, d)),
    valueOf0(bf, d)
  )
})