Track the evolution of a function along merging

For a given ARI merging, compute the evolution on the function f

functionTracking(merger, f, p = 1, n_steps = NULL, ...)

Arguments

merger	the result from having run Dune on the dataset
f	the function used. It must takes as input a clustering matrix and return a value
p	A value between 0 and 1. We stop when the mean ARI has improved by p of the final total improvement. Default to 1 (i.e running the full merging).
n_steps	Alternatively, you can specifiy the number of merging steps to do before stopping.
...	additional arguments passed to f

Value

a vector of length the number of merges

Examples

# Return the number of clusters for the fourth cluster label
data("clusMat", package = "Dune")
merger <- Dune(clusMat = clusMat)
f <- function(clusMat, i) dplyr::n_distinct(clusMat[, i])
functionTracking(merger, f, i = 4)
#> [1] 12 12 12 11 11 11 10 10