pkbc_validation {QuadratiK} | R Documentation |
Validation of Poisson kernel-based clustering results
Description
Method for objects of class pkbc
which computes evaluation measures
for clustering results.
Usage
pkbc_validation(object, true_label = NULL, h = 1.5)
Arguments
object |
Object of class |
true_label |
factor or vector of true membership to clusters (if available). It must have the same length of final memberships. |
h |
Tuning parameter of the k-sample test. (default: 1.5) |
Details
The following evaluation measures are computed: In-Group Proportion. If true label are provided, ARI, Average Silhouette Width, Macro-Precision and Macro-Recall are computed.
Value
List with the following components:
-
metrics
Table of computed evaluation measures. -
IGP
List of in-group proportions for each value of number of clusters specified.
References
Kapp, A.V., Tibshirani, R. (2007) "Are clusters found in one dataset present in another dataset?", Biostatistics, 8(1), 9–31, https://doi.org/10.1093/biostatistics/kxj029
Rousseeuw, P.J. (1987) Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20, 53–65.
Examples
#We generate three samples of 100 observations from 3-dimensional
#Poisson kernel-based densities with rho=0.8 and different mean directions
size<-20
groups<-c(rep(1, size), rep(2, size),rep(3,size))
rho<-0.8
set.seed(081423)
data1<-rpkb(size, c(1,0,0),rho,method='rejvmf')
data2<-rpkb(size, c(0,1,0),rho,method='rejvmf')
data3<-rpkb(size, c(1,0,0),rho,method='rejvmf')
data<-rbind(data1$x,data2$x, data3$x)
#Perform the clustering algorithm
pkbc_res<- pkbc(data, 2:4)
pkbc_validation(pkbc_res)