A hidden process regression model for functional data description. Application to curve discrimination

Faicel Chamroukhi (1, 2) , Allou Samé (2) , Gérard Govaert (1) , Patrice Aknin (2)

1 Heudiasyc - Heuristique et Diagnostic des Systèmes Complexes [Compiègne]
2 INRETS/LTN - Laboratoire des Technologies Nouvelles

A new approach for functional data description is proposed in this paper. It consists of a regression model wit a discrete hidden logistic process which is adapted for modeling curves with abrupt or smooth regime changes.The model parameters are estimated in a maximum likelihood framework through a dedicated expectation maximization (EM) algorithm.From the proposed generative model, a curve discrimination rule is derived using the maximum a posteriori rule. The proposed model is evaluated using simulated curves and real world curves acquired during railways witch operations, by performing comparisons with the piecewise regression approach in terms of curv emodeling and classification.

Domaines: Statistiques [math.ST]; Théorie [stat.TH]

Faicel Chamroukhi, Allou Samé, Gérard Govaert, Patrice Aknin. A hidden process regression model for functional data description. Application to curve discrimination. Neurocomputing, 2010, 73 (7-9), pp.1210-1221. ⟨10.1016/j.neucom.2009.12.023⟩. ⟨hal-00485163⟩