Conference on non stationarity

Conference organized by the AGM laboratory on June 4-5, 2018 at the MIR, Neuville-sur-Oise and on June 6, 2018 in the salle des conférences, Chênes 1

The techniques of time series are widely used for the prediction of random phenomena. Other related questions such as for example calibration, resampling or risk management are other important questions…

The stationarity is usually assumed because of a simple technical feature ; it makes possible to use the ergodic theorem in order to derive consistent estimations.

The aim of the conférence is to consider models which dont fit this condition. The considered models should fit with considered data sets, and beyond the local stationarity  Les modèles suggérés devront être en accord avec les données considérées et outre les conditions de stationnarité locales (Dahlhaus) we aim at extending the the field of possible models involving periodicities, exogenous data or shape constraints as monotonicity. Applications will support the considered mathematical models.

Coordinator : paul.doukhan@u-cergy.fr

Read more