gama (version 1.0.3)

gama.how.many.k: Estimates the optimal number of partitions.

Description

This function estimates the best k value for the number of partitions the dataset should be segmented.

Usage

gama.how.many.k(dataset = NULL, method = "minimal")

Arguments

dataset

the original dataset used for clustering.

method

the method used to estimate the number of partitions. If 'minimal' is used, the function will perform estimation based on finding the 'elbow' in the Within-cluster Sum of Squares Error graphic. It uses a second derivative approximation, in order to suggest k. If 'broad' is used, the function will proceed an estimation by majority voting of 24 indices, by using the NbClust package.

References

Malika Charrad, Nadia Ghazzali, Veronique Boiteau, Azam Niknafs (2014). NbClust: An R Package for Determining the Relevant Number of Clusters in a Data Set. Journal of Statistical Software, 61(6), 1-36. URL http://www.jstatsoft.org/v61/i06/.

See Also

gama.

Examples

# NOT RUN {
# loads data about CPU execution metrics of a distributed
# version of Alternating Least Squares (ALS) algorithm
library(gama)
data(cpu.als)

# call estimation by using minimal method (Elbow graphic)
k <- gama.how.many.k (cpu.als)
print(k)

# call estimation by using broad method (NbClust)
k <- gama.how.many.k (cpu.als, method = 'broad')
print(k)

# }