kmeans::RefineMiniBatchOptions Struct Reference

Options for RefineMiniBatch construction. More...

#include <RefineMiniBatch.hpp>

Public Attributes
int	max_iterations = 100
int	batch_size = 500
double	max_change_proportion = 0.01
int	convergence_history = 10
uint64_t	seed = 1234567890u
int	num_threads = 1

Detailed Description

Options for RefineMiniBatch construction.

Member Data Documentation

batch_size

int kmeans::RefineMiniBatchOptions::batch_size = 500

Number of observations in the mini-batch. Larger numbers improve quality at the cost of computational time and memory.

convergence_history

int kmeans::RefineMiniBatchOptions::convergence_history = 10

Number of iterations to remember when checking for convergence (i.e., \(h\) in the documentation for RefineMiniBatch). Larger values improve the quality of the result at the cost of computational time.

max_change_proportion

double kmeans::RefineMiniBatchOptions::max_change_proportion = 0.01

Maximum proportion of observations in each cluster that can be reassigned when checking for convergence (i.e., \(p\) in the documentation for RefineMiniBatch). Lower values improve the quality of the result at the cost of computational time.

max_iterations

int kmeans::RefineMiniBatchOptions::max_iterations = 100

Maximum number of iterations. More iterations increase the opportunity for convergence at the cost of more computational time.

num_threads

int kmeans::RefineMiniBatchOptions::num_threads = 1

Number of threads to use. The parallelization scheme is defined by parallelize().

seed

uint64_t kmeans::RefineMiniBatchOptions::seed = 1234567890u

Seed to use for the PRNG when sampling observations to use in each mini-batch.

---

The documentation for this struct was generated from the following file:

• kmeans/RefineMiniBatch.hpp