make_experiment: Run a core clustering experiment on a dataset.

In bwrc/corecluster-r: Core clustering of datasets.

Description

Run a core clustering experiment on a dataset.

Usage

make_experiment(dataset, clustering_func = NULL, sampling_func = NULL,
  alpha = 0.1, n_iter = 1000, method = "bootstrap", save = NULL,
  seed = 42, using_trim = FALSE)

Arguments

dataset	A dataset: a list with (at least) the fields data (the data as a matrix or data frame, depending on the clustering function) and classes (the true, known classes of the dataset).
clustering_func	A clustering function to use, which given data returns cluster indices.
sampling_func	A sampling function to use. Default is NULL. Only used when method is distribution.
alpha	The confidenence level. Default 0.1
n_iter	The number of iterations to use. The default is 1000.
method	The method to use. String. Either bootstrap to use bootstrapping or distribution to calculate the exact co-occurrence probabilities.
save	The full path to a file (rds-file) where to save the data.
seed	Random seed. Default is NULL to not set the random seed.
using_trim	Boolen. Does the clustering method use trimming?

Value

A list structure:

data

The data

alpha

The confidence level

iterations

The number of iterations

cluster

The initial clustering

cluster_core

The core clustering

mat

The co-occurrence probability matrix

amat

The adjacency matrix

using_trim

Whether or not the clustering method is using trimming

weak_cluster

The index used to describe the weak clusters

classes_true

The known true classes of the data items

clustering_function

The used clustering function

time_start

The timestamp when the experiment was started

time_stop

The timestamp when the experiment was ready

bwrc/corecluster-r documentation built on May 13, 2019, 9:12 a.m.