tests/testthat/test-change.R

In fChange: Functional Change Point Detection and Analysis

## SINGLE CHANGES
test_that("Characteristic change", {
  if (Sys.info()["sysname"] == "Linux") {
    expect_false(FALSE)
    return()
  }
  ## No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(
    X = dat, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.835)

  ## Change
  set.seed(123)
  tmp <- fchange(electricity,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0)


  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################
  ## No Change
  set.seed(123)
  tmp <- fchange(dat,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.755)

  set.seed(123)
  tmp <- fchange(
    X = dat, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "welch",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.810)
  expect_error(change(dat,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "welch",
    M = 200
  ))

  set.seed(123)
  tmp <- fchange(
    X = dat, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "resample",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.880)
  tmp <- fchange(dat,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "resample",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.700)


  ## Change
  set.seed(123)
  tmp <- fchange(
    X = electricity, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0)
  expect_equal(tmp$pvalue, 0)

  set.seed(123)
  tmp <- fchange(
    X = electricity, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "welch",
    M = 200
  )
  # expect_equal(tmp$pvalue, 0)
  expect_error(change(electricity,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "welch",
    M = 200
  ))

  set.seed(123)
  tmp <- fchange(
    X = electricity, method = "characteristic", h = 0,
    statistic = "Tn", type = "Single",
    critical = "resample",
    M = 200
  )
  expect_equal(tmp$pvalue, 0)
  tmp <- fchange(electricity,
    method = "characteristic", h = 0,
    statistic = "Mn", type = "Single",
    critical = "resample",
    M = 200
  )
  expect_equal(tmp$pvalue, 0)
})


test_that("Mean change", {
  # Simulation - No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "mean",
    statistic = "Tn", type = "Single",
    critical = "simulation",
    M = 1000, K = bartlett_kernel
  )
  expect_equal(tmp$pvalue, 0.511)

  tmp <- fchange(electricity,
    method = "mean",
    statistic = "Mn", type = "Single",
    critical = "resample",
    M = 1000, K = bartlett_kernel
  )
  expect_equal(tmp$pvalue, 0)

  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  # Simulation - No Change
  set.seed(123)
  tmp <- fchange(dat,
    method = "mean",
    statistic = "Mn", type = "Single",
    critical = "simulation",
    M = 1000, K = bartlett_kernel
  )
  expect_equal(tmp$pvalue, 0.579)

  # resample - Change
  set.seed(1234)
  tmp <- fchange(electricity,
    method = "mean",
    statistic = "Tn", type = "Single",
    critical = "resample",
    M = 1000, K = bartlett_kernel
  )
  expect_equal(tmp$pvalue, 0)
})


test_that("Robust change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "robustmean",
    statistic = "Tn", type = "Single",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.493)

  # Change
  set.seed(134)
  dat_change <- dat
  dat_change$data[, 50:100] <- dat_change$data[, 50:100] + 5

  tmp <- fchange(dat_change,
    method = "robustmean",
    statistic = "Mn", type = "Single",
    critical = "resample",
    M = 500
  )
  expect_equal(tmp$pvalue, 0.014)

  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################
  # No Change
  set.seed(123)
  tmp <- fchange(dat,
    method = "robustmean",
    statistic = "Mn", type = "Single",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.778)

  # Change
  set.seed(213)
  dat_change <- dat
  dat_change$data[, 50:100] <- dat_change$data[, 50:100] + 5

  tmp <- fchange(dat_change,
    method = "robustmean",
    statistic = "Tn", type = "Single",
    critical = "resample",
    M = 500
  )
  expect_equal(tmp$pvalue, 0.008)
})


test_that("Eigen change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "eigenjoint", h = 0, eigen_number = 1,
    statistic = "Tn",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.240)

  tmp <- fchange(electricity,
    method = "eigensingle", h = 0, eigen_number = 1,
    statistic = "Mn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0)

  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################
  ## No Change
  set.seed(123)
  tmp <- fchange(dat,
    method = "eigensingle", h = 0, eigen_number = 1,
    statistic = "Tn",
    critical = "simulation",
    M = 200
  )
  expect_equal(tmp$pvalue, 0.210)

  set.seed(123)
  tmp <- fchange(dat,
    method = "eigenjoint", h = 0, eigen_number = 1,
    statistic = "Mn",
    critical = "resample",
    M = 5
  )
  expect_equal(tmp$pvalue, 0.4)
  tmp <- fchange(dat,
    method = "eigensingle", h = 0, eigen_number = 1,
    statistic = "Mn",
    critical = "resample",
    M = 5
  )
  expect_equal(tmp$pvalue, 0)

  ## Change
  set.seed(123)
  tmp <- fchange(electricity,
    method = "eigenjoint", h = 0, eigen_number = 1,
    statistic = "Tn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0)
  tmp <- fchange(electricity,
    method = "eigensingle", h = 0, eigen_number = 1,
    statistic = "Tn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0)

  tmp <- fchange(electricity,
    method = "eigenjoint", h = 0, eigen_number = 1,
    statistic = "Mn",
    critical = "resample",
    M = 2
  )
  expect_equal(tmp$pvalue, 0)
  tmp <- fchange(electricity,
    method = "eigensingle", h = 0, eigen_number = 1,
    statistic = "Mn",
    critical = "resample",
    M = 10
  )
  expect_equal(tmp$pvalue, 0)
})


test_that("Trace change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "trace",
    statistic = "Tn", type = "Single",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.292)


  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################
  # No change
  tmp <- fchange(dat,
    method = "trace",
    statistic = "Mn", type = "Single",
    critical = "resample",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.150)

  # Change
  tmp <- fchange(electricity,
    method = "trace",
    statistic = "Tn", type = "Single",
    critical = "resample",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.016)

  tmp <- fchange(electricity,
    method = "trace",
    statistic = "Mn", type = "Single",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.011)
})


test_that("Covariance change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "covariance",
    statistic = "Tn",
    critical = "simulation",
    M = 1000, cov.res = 20
  )
  expect_equal(tmp$pvalue, 0.051)


  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  set.seed(123)
  tmp <- fchange(dat,
    method = "covariance",
    statistic = "Tn",
    critical = "resample",
    M = 10, cov.res = 5
  )
  expect_equal(tmp$pvalue, 0.2)

  expect_error(change(dat,
    method = "covariance",
    statistic = "Mn",
    critical = "simulation",
    M = 1000
  ))

  # Change
  set.seed(123)
  tmp <- fchange(electricity,
    method = "covariance",
    statistic = "Tn",
    critical = "simulation",
    M = 1000, cov.res = 15
  )
  expect_equal(tmp$pvalue, 0)
})


test_that("PCA-Mean change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  tmp <- fchange(dat,
    method = "projmean",
    statistic = "Tn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.210)


  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  # No Change
  tmp <- fchange(dat,
    method = "projmean",
    statistic = "Tn",
    critical = "resample",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.067)

  tmp <- fchange(dat,
    method = "projmean",
    statistic = "Mn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.244)
  tmp <- fchange(dat,
    method = "projmean",
    statistic = "Mn",
    critical = "resample",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0.135)

  # Change
  set.seed(123)
  tmp <- fchange(electricity,
    method = "projmean",
    statistic = "Tn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0)
  tmp <- fchange(electricity,
    method = "projmean",
    statistic = "Mn",
    critical = "simulation",
    M = 1000
  )
  expect_equal(tmp$pvalue, 0)
})


test_that("PCA-Distribution change", {
  # No Change
  set.seed(123)
  dat <- generate_brownian_bridge(50)
  tmp <- fchange(dat,
    method = "projdistribution",
    statistic = "Tn",
    critical = "resample",
    M = 100
  )
  expect_equal(tmp$pvalue, 0.530)
  tmp <- fchange(dat,
    method = "projdistribution",
    statistic = "Mn",
    critical = "resample",
    M = 100
  )
  expect_equal(tmp$pvalue, 0.53)


  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  # No Change
  set.seed(123)
  tmp <- fchange(dat,
    method = "projdistribution",
    statistic = "Mn",
    critical = "resample",
    M = 100
  )
  expect_equal(tmp$pvalue, 0.610)

  # Change
  dat$data[, 25:50] <- dat$data[, 25:50] + 5
  tmp <- fchange(dat,
    method = "projdistribution",
    statistic = "Tn",
    critical = "resample",
    M = 100
  )
  expect_equal(tmp$pvalue, 0)


  # Error
  expect_error(change(dat,
    method = "projdistribution",
    statistic = "Mn",
    critical = "simulation",
    M = 100
  ))
})


## ELBOW PLOTS
test_that("Check Elbow Plots", {
  ## No Change
  set.seed(123)
  dat <- generate_brownian_bridge(100)
  res <- fchange(
    X = dat,
    method = "characteristic",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)
  expect_true(ggplot2::is_ggplot(res$suggestion$plot))
  expect_true(ggplot2::is_ggplot(res$plots$variability))
  expect_true(ggplot2::is_ggplot(res$plots$explained))
  expect_true(ggplot2::is_ggplot(res$plots$improvement))
  expect_equal(round(sum(res$information[, 2]), 4), 35.6545)

  ## Change
  set.seed(123)
  dat_change <- generate_brownian_bridge(100)
  dat_change$data[, 51:100] <- dat_change$data[, 51:100] + 5
  res <- fchange(
    X = dat_change,
    method = "characteristic",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)
  expect_true(ggplot2::is_ggplot(res$suggestion$plot))
  expect_true(ggplot2::is_ggplot(res$plots$variability))
  expect_true(ggplot2::is_ggplot(res$plots$explained))
  expect_true(ggplot2::is_ggplot(res$plots$improvement))
  expect_equal(round(sum(res$information[, 2]), 4), 224.252)

  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################
  ## No Changes

  set.seed(123)
  res <- fchange(
    X = dat,
    method = "mean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "robustmean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "eigenjoint",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "eigensingle",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "trace",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "covariance",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "projmean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat,
    method = "projdistribution",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)



  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  ## Changes
  set.seed(123)
  res <- fchange(
    X = dat_change,
    method = "mean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)

  res <- fchange(
    X = dat_change,
    method = "robustmean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)

  res <- fchange(
    X = dat_change,
    method = "eigenjoint",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, c(52, 49, 50))

  res <- fchange(
    X = dat_change,
    method = "eigensingle",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, c(52, 49, 50))

  res <- fchange(
    X = dat_change,
    method = "trace",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  res <- fchange(
    X = dat_change,
    method = "covariance",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, c(39, 50))

  res <- fchange(
    X = dat_change,
    method = "projmean",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)

  res <- fchange(
    X = dat_change,
    method = "projdistribution",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "L2", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)

  ## Errors
  set.seed(1234)
  dat <- generate_brownian_bridge(100)
  res <- fchange(
    X = dat,
    method = "characteristic",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "trace", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, NA)

  dat$data[, 51:100] <- dat$data[, 51:100] + 5
  res <- fchange(
    X = dat,
    method = "characteristic",
    type = "Elbow",
    max_changes = min(ncol(dat), 20),
    eigen_number = 2, h = 1,
    W = space_measuring_functions(X = dat, M = 20, space = "BM"),
    K = bartlett_kernel,
    weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    errors = "trace", recommendation_change_points = 2,
    recommendation_improvement = 0.15
  )
  expect_equal(res$suggestion$changes, 50)
})


## Binary Segmentation
test_that("Binary Segmentation", {
  set.seed(12345)
  X <- generate_brownian_bridge(150)
  X[, 51:150] <- X[, 51:150] + 3
  X[, 101:150] <- X[, 101:150] * 2

  res <- fchange(X,
    method = "characteristic",
    statistic = "Tn",
    critical = "welch",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(50, 99))
  expect_equal(res$pvalue, c(0.000, 0.000))

  #################
  ## Only run below locally
  testthat::skip_on_cran()
  #################

  res <- fchange(X,
    method = "mean",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(50, 66, 100))
  expect_equal(res$pvalue, c(0.001, 0, 0))


  res <- fchange(X,
    method = "robustmean",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(50, 66, 100))
  expect_equal(res$pvalue, c(0, 0, 0))

  res <- fchange(X,
    method = "eigenjoint",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(102))
  expect_equal(res$pvalue, c(0))

  res <- fchange(X,
    method = "eigensingle",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(102))
  expect_equal(res$pvalue, c(0.003))

  res <- fchange(X,
    method = "trace",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, NULL)
  expect_equal(res$pvalue, NULL)

  res <- fchange(X,
    method = "projmean",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(50, 66, 100))
  expect_equal(res$pvalue, c(0, 0, 0))


  set.seed(123)
  X <- generate_brownian_bridge(50)
  X$data[, 26:50] <- X$data[, 26:50] + 5

  res <- fchange(X,
    method = "projdistribution",
    statistic = "Tn",
    critical = "resample",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, c(25))
  expect_equal(res$pvalue, c(0))

  res <- fchange(X,
    method = "covariance",
    statistic = "Tn",
    critical = "simulation",
    type = "segmentation",
    resample_blocks = "separate", replace = TRUE,
    blocksize = 1,
    eigen_number = 3, h = 3,
    M = 1000, J = 50,
    W = space_measuring_functions(X = X, M = 20, space = "BM"),
    K = bartlett_kernel,
    alpha = 0.05, cov.res = 30, weighting = 1 / 4, TVE = 0.95,
    trim_function = function(X) {
      0
    },
    silent.binary = TRUE
  )
  expect_equal(res$location, NULL)
  expect_equal(res$pvalue, NULL)
})