SegAnnot: an R package for fast segmentation of annotated piecewise constant signals

Toby Dylan Hocking (1, 2, 3) , Guillem Rigaill (4)

1 SIERRA - Statistical Machine Learning and Parsimony
2 CBIO - Centre de Bioinformatique
3 Cancer et génome: Bioinformatique, biostatistiques et épidémiologie d'un système complexe
4 MIA-Paris - Mathématiques et Informatique Appliquées

We describe and propose an implementation of a dynamic programming algorithm for the segmentation of annotated piecewise constant signals. The algorithm is exact in the sense that it recovers the best possible segmentation w.r.t. the quadratic loss that agrees with the annotations.

Domaines: Machine Learning [stat.ML]; Bio-informatique [q-bio.QM]; Bio-Informatique, Biologie Systémique [q-bio.QM]; Optimisation et contrôle [math.OC]

Toby Dylan Hocking, Guillem Rigaill. SegAnnot: an R package for fast segmentation of annotated piecewise constant signals. 2012. ⟨hal-00759129⟩