Categorical functional data analysis. The cfda R package
Analyse des données fonctionnelles catégorielles. Le package R cfda
Résumé
We present how to take into account categorical functional data, represented by the trajectories of a jump process with a continuous time and a finite set of states. As an extension of multiple correspondence analysis to an infinite set of variables,
the optimal codings of the states over time are approximated on an arbitrary finite function basis. This allows dimension reduction, representation optimization and visualisation of data in smaller dimensional spaces. We have implemented this methodology in the R
package cfda available on the CRAN, we will present in this communication how it can be implemented on real data in a clustering framework.
Origine | Fichiers produits par l'(les) auteur(s) |
---|