Package for Brik, Fabrik and Fdebrik Algorithms to Initialise
Kmeans
Description
Implementation of the BRIk, FABRIk and FDEBRIk algorithms
to initialise k-means. These methods are intended for the
clustering of multivariate and functional data, respectively.
They make use of the Modified Band Depth and bootstrap to
identify appropriate initial seeds for k-means, which are
proven to be better options than many techniques in the
literature. Torrente and Romo (2021)
It makes use of the functions kma and kma.similarity, from the
archived package fdakma, by Alice Parodi et al.