B0266
Title: Robust inference for change points in piecewise polynomials of general degrees
Authors: Shakeel Gavioli-Akilagun - London School of Economics (United Kingdom) [presenting]
Piotr Fryzlewicz - London School of Economics (United Kingdom)
Abstract: Multiple change-point detection has become popular with the routine collection of complex non-stationary time series. An equally important but comparatively neglected question concerns quantifying the level of uncertainty around each putative changepoint. Though a handful of procedures exist in the literature, most all make assumptions on the density of the contaminating noise which are impossible to verify in practice. We present a procedure that, under minimal assumptions, returns localized regions of a data sequence that must contain a changepoint at some global significance level chosen by the user. Our procedure is computationally efficient, applicable to change points in higher-order polynomials, and moreover, all results are fully non-asymptotic. We will discuss some appealing theoretical properties of our procedure, and show its good practical performance on real and simulated data.