README.md

In Ltochon/CLARA.seq: CLARA Clustering

CLARA.seq

R package for sequences clustering This package works with the TraMineR library which provides the LCS method for distance calculations.

Installation

To use this package on R, this following lines are required :

install.packages("devtools")
library(devtools)
install_github("Ltochon/CLARA.seq")

If the vignette are not refreshed and you can't see the CLARA.seq's one, use this line :

devtools::install(build_vignettes = TRUE)

Algorithms

This package contains 3 differents algorithms - CLARA - CLARANS - CLARA-FUZZY

And a quality index - Davies-Bouldin Index

Purpose of this package

TraMineR's package has a limitation of the number of sequences used (~46'300). To counter this limit, different algorithm has been implemented to use subsets of the entire dataset to extract the best clustering for the big dataset. At this moment, a dataset with 227'000 sequences has been tested and perfectly clustered with the CLARA algorithm.

WARNINGS CLARA algorithm is the most efficient one. CLARANS and CLARA-FUZZY are still in test phase.

Documentation

The complete documentation is available in the folder Documentation - User guide - Package Documentation

Ltochon/CLARA.seq documentation built on Dec. 17, 2021, 1:12 a.m.