Elastic Hierarchical Clustering

Perform agglomerative hierarchical clustering on functional data using pairwise elastic (Fisher-Rao) distances. Supports single, complete, and average linkage.

Usage

elastic.hclust(
  fdataobj,
  method = c("complete", "single", "average"),
  lambda = 0
)

Arguments

fdataobj

An object of class fdata.

method

Character: linkage method. One of "complete" (default), "single", or "average".

lambda

Regularization parameter for elastic alignment (default 0).

Value

An object of class elastic.hclust with components:

merges

A 3-column matrix of merge steps (i, j, distance), with 1-indexed cluster labels

distance.matrix

The pairwise elastic distance matrix

method

The linkage method used

fdataobj

Original fdata object

call

The matched call

References

Srivastava, A. and Klassen, E. (2016). Functional and Shape Data Analysis. Springer.

See also

elastic.cutree for cutting the dendrogram, elastic.kmeans for k-means clustering, elastic.distance for the distance metric

Examples

# \donttest{
set.seed(1)
t <- seq(0, 1, length.out = 30)
X <- matrix(0, 10, 30)
for (i in 1:5)  X[i, ] <- sin(2 * pi * t) + rnorm(30, 0, 0.1)
for (i in 6:10) X[i, ] <- cos(2 * pi * t) + rnorm(30, 0, 0.1)
fd <- fdata(X, argvals = t)
hc <- elastic.hclust(fd, method = "complete")
hc
#> Elastic Hierarchical Clustering
#>   Curves: 10 x 30 grid points
#>   Method: complete 
#>   Merge steps: 9 
# }