inst/tinytest/test-objective-value.R

In anticlust: Subset Partitioning via Anticlustering

library("anticlust")

# objective value for variance criterion is computed correctly
for (m in 1:4) {
  m_features <- m
  ## vary number of anticlusters
  for (p in 2:5) {
    p_anticlusters <- p
    n_elements <- p * 3 # n-elements must be multiplier of p
    features <- matrix(rnorm(n_elements * m_features), ncol = m_features)
    cl <- kmeans(features, p_anticlusters)
    obj_kmeans <- cl$tot.withinss
    obj_mine  <- variance_objective(features, cl$cluster)
    expect_equal(obj_kmeans, obj_mine)
  }
}

# objective value for distance criterion is computed correctly
conditions <- expand.grid(m = 1:4, p = 2:3)
for (k in 1:nrow(conditions)) {
  m_features <- conditions[k, "m"]
  p_anticlusters <- conditions[k, "p"]
  n_elements <- p_anticlusters * 3 # n must be multiplier of p
  features <- matrix(rnorm(n_elements * m_features), ncol = m_features)
  distances <- as.matrix(dist(features))
  ilp <- anticlust:::anticlustering_ilp(distances, p_anticlusters)
  solution <- anticlust:::solve_ilp(ilp, "min")
  anticlusters <- anticlust:::ilp_to_groups(solution, n_elements)
  expect_equal(solution$obj, anticlust:::diversity_objective_(anticlusters, features))
}