This vignette illustrates the basic usage of the PAC package for R.
For simplicity, we will generate the synthetic data from a mixture of gaussian. The details of the construction are unimportant, but are provided below for completeness.
# Problem parameters
n = 5e3 # number of observations
p = 1 # number of dimensions
K = 3 # number of clusters
w = rep(1,K)/K # component weights
mu <- c(0,2,4) # component means
sd <- rep(1,K)/K # component standard deviations
# Generate K mixture of gaussian
g <- sample(1:K,prob=w,size=n,replace=TRUE) # ground truth for clustering
X <- as.matrix(rnorm(n=n,mean=mu[g],sd=sd[g]))
The main entry point for the PAC package is the function PAC. To begin, we call PAC with all of the default options.
library(PAC)
y <- PAC(X, K)
## Input Data: 5000 by 1
## Partition method: Discrepancy based partition
## Maximum level: 40
## partition completed
## [1] "Initial Clustering..."
## [1] "Merging..."
We can evaluate the accuracy of the clustering result by computing the fmeasure between it and the ground truth: of the default options.
print(fmeasure(g,y))
## [1] 0.9940302
We now show how to change the default value of PAC's has two optional arguments: the maximum level of the partition(“maxlevel”) and the partition method(“method”).
For example, one can use a larger number of partitions (the default is 40):
y.deep <- PAC(X, K, maxlevel = 100)
## Input Data: 5000 by 1
## Partition method: Discrepancy based partition
## Maximum level: 100
## partition completed
## [1] "Initial Clustering..."
## [1] "Merging..."
print(fmeasure(g,y.deep))
## [1] 0.9801951
Or, one can change the partition method to be bsp with limited lookahead See [@lu2013multivariate], [@wong2010optional]. The default is dsp [@yang2014density].
y.ll <- PAC(X, K, method = 'll')
## Input Data: 5000 by 1
## Partition method: BSP with limited-lookahead
## Maximum level: 40
## partition completed
## [1] "Initial Clustering..."
## [1] "Merging..."
print(fmeasure(g,y.ll))
## [1] 0.9956008