Comparison of three algorithms for parametric change-point detection

Jean-Marc Bardet (1) , Faure Cynthia (1) , Jérôme Lacaille (2) , Madalina Olteanu (1)

1 SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne)
2 SNECMA Villaroche [Moissy-Cramayel]

Numerous sensors placed on aircraft engines capture a consid- erable amount of data during tests or flights. In order to detect potential crucial changes of characteristic features, it is relevant to develop powerful statistical algorithms. This manuscript aims at detecting change-points, in an off-line framework, in piecewise-linear models and with an unknown number of change-points. In this context, three recent algorithms are con- sidered, implemented and compared on simulated and real data.

Mots clés: change-point detection; time series; health monitoring

Domaines: Statistiques [math.ST]

Jean-Marc Bardet, Faure Cynthia, Jérôme Lacaille, Madalina Olteanu. Comparison of three algorithms for parametric change-point detection . European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2016), 2016, Bruges, Belgium. ⟨hal-01519702⟩