clustercheck: Checking cluster accuracy

In CAMCOS/camcos2017: CAMCOS Spring 2017: Efficient and Scalable Spectral Clustering for Document Clustering.

Description

In the results table, true clusters are the rows, guessed clusters are the columns

Usage

clustercheck(cluster, Labels, k)

Arguments

cluster	Integer vector. Labels from clustering
Labels	Integer vector. Ground truth labels.
k	Integer. Number of clusters.

Value

List of 3. Confusion matrix, label mapping, accuracy.

Examples

# Example of a results matrix that FAILS clustercheck
# Cluster assignment of "true cluster #3" is split too evenly
results <- matrix(c(15,31,4,452,5,0,1,384,1,20,1,3,34,141,56,53,289,357,0,
  2,282,6,160,5,527,32,48,51,9,6,4,2,203,0,0,5), nrow = 6)

CAMCOS/camcos2017 documentation built on May 6, 2019, 9:23 a.m.