Exploratory Analysis of Functional Data via Clustering and Optimal Segmentation

Georges Hébrail (1, 2) , Bernard Hugueney (3) , Yves Lechevallier (4) , Fabrice Rossi (1, 2)

1 BILab - Laboratoire de Business Intelligence Télécom-ParisTech/EDF
2 LTCI - Laboratoire Traitement et Communication de l'Information
3 LAMSADE - Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision
4 AxIS - Usage-centered design, analysis and improvement of information systems

We propose in this paper an exploratory analysis algorithm for functional data. The method partitions a set of functions into $K$ clusters and represents each cluster by a simple prototype (e.g., piecewise constant). The total number of segments in the prototypes, $P$, is chosen by the user and optimally distributed among the clusters via two dynamic programming algorithms. The practical relevance of the method is shown on two real world datasets.

Mots clés: Functional Data; Multiple time series; Exploratory analysis; Clustering; Segmentation; Dynamic programming

Domaines: Apprentissage [cs.LG]; Autres [stat.ML]

Georges Hébrail, Bernard Hugueney, Yves Lechevallier, Fabrice Rossi. Exploratory Analysis of Functional Data via Clustering and Optimal Segmentation. Neurocomputing, 2010, 73 (7-9), pp.Pages 1125-1141. ⟨10.1016/j.neucom.2009.11.022⟩. ⟨hal-00515908⟩