kmeans
k-means clustering in C++
|
Options for RefineHartiganWong
.
More...
#include <RefineHartiganWong.hpp>
Public Attributes | |
int | max_iterations = 10 |
int | max_quick_transfer_iterations = 50 |
bool | quit_on_quick_transfer_convergence_failure = false |
int | num_threads = 1 |
Options for RefineHartiganWong
.
int kmeans::RefineHartiganWongOptions::max_iterations = 10 |
Maximum number of optimal transfer iterations. More iterations increase the opportunity for convergence at the cost of more compute time.
int kmeans::RefineHartiganWongOptions::max_quick_transfer_iterations = 50 |
Maximum number of quick transfer iterations. More iterations increase the opportunity for convergence at the cost of more compute time.
int kmeans::RefineHartiganWongOptions::num_threads = 1 |
Number of threads to use. The parallelization scheme is defined by parallelize()
.
bool kmeans::RefineHartiganWongOptions::quit_on_quick_transfer_convergence_failure = false |
Whether to quit early when the number of quick transfer iterations exceeds RefineHartiganWongOptions::max_quick_tranfer_iterations
. Setting this to true recovers the default behavior of R's kmeans()
implementation. If false, the algorithm will ignore any convergence failures during the quick transfer step and procced to the next optimal transfer iteration. This provides another opportunity to improve the clustering (and potentially achieve convergence) at the cost of more compute time.