TSclust-package: Package for Time Series Clustering.

In TSclust: Time Series Clustering Utilities

Description

This package contains several measures of dissimilarity between time series, some examples of time series datasets, specific clustering algorithms, and dimension reduction algorithms. dissimilarities begin with diss.*, and a wrapper function diss is available. Cluster evaluation methods include cluster.evaluation and loo1nn.cv. A clustering algorithm based on pairwise p-values is implemented in pvalues.clust. The package should be used along with other existing clustering packages and function such as hclust, packages cluster, ...

Author(s)

Pablo Montero Manso, José Antonio Vilar.

References

Montero, P and Vilar, J.A. (2014) TSclust: An R Package for Time Series Clustering. Journal of Statistical Software, 62(1), 1-43. http://www.jstatsoft.org/v62/i01/.

See Also

diss

Examples

#the available dissimilarities can be found in the diss help, page (?diss)
#and their individual pages from there.

### The most common use case begins with a set of time series we want to cluster.
### This package includes several example datasets.
### 
data(interest.rates)
###transformation of the interest rates
trans.inter.rates <- log(interest.rates[2:215,]) - log(interest.rates[1:214,])

##use the dist function of the proxy package to easily create the dist object
#applying ACF with geometric decaying to each pair of time series
tsdist <- diss( t(trans.inter.rates) , "ACF", p=0.05)

names(tsdist) <- colnames(interest.rates)

#perform hierachical clustering to the dist object
hc <- hclust(tsdist)

#show the results
plot(hc)

mahdist <- diss( t(trans.inter.rates) , "AR.MAH", p=0.05)$p_value

pvalues.clust(mahdist, 0.05)

TSclust documentation built on July 23, 2020, 1:07 a.m.