Consistent change-point detection with kernels

Abstract

In this paper we study the kernel change-point algorithm (KCP) proposed by Arlot, Celisse and Harchaoui [4], which aims at locating an unknown number of change-points in the distribution of a sequence of independent data taking values in an arbitrary set. The change-points are selected by model selection with a penalized kernel empirical criterion. We provide a non-asymptotic result showing that, with high probability, the KCP procedure retrieves the correct number of change-points, provided that the constant in the penalty is well-chosen; in addition, KCP estimates the change-points location at the optimal rate. As a consequence, when using a characteristic kernel, KCP detects all kinds of change in the distribution (not only changes in the mean or the variance), and it is able to do so for complex structured data (not necessarily in R d ). Most of the analysis is conducted assuming that the kernel is bounded; part of the results can be extended when we only assume a finite second-order moment.MSC 2010 subject classifications: Primary 62M10; secondary 62G20.

New efficient algorithms for multiple change-point detection with reproducing kernels

“…This is the usual non-asymptotic optimality result for model 6 selection procedures [12]. This procedure has also been proved to provide consistent estimates of the change-points [23]. However, from a computational point of view, the first step (i.e.…”

Robust semi-parametric multiple change-points detection

“…where κ is obtained from a slope heuristic procedure and the function ψ is given in (19). There is only one general relation (11) between those estimators, i.e.…”

“…[17] focuses on a cross-validation selection method and [18] is placed in the context of heteroscedastic data. Other nonparametric approaches exist, for example [19] proposes kernel estimators, also see [10] in a different framework. [20] investigates the fused LASSO procedure and showed some difficulties.…”

Selective review of offline change point detection methods

“…From a theoretical point of view, asymptotic consistency of the change point estimates has been demonstrated for both a known and unknown number of change points in the recent work of [102]. This result, as well as an important oracle inequality on the sum of cost V (T ) [103], also holds in a non-asymptotic setting.…”