Change detection via affine and quadratic detectors (Q1689004)

The goal of the paper is to develop a specific application of the convex optimization based hypothesis testing techniques developed in [the second and the last author, Electron. J. Stat. 10, No. 2, 2204--2242 (2016; Zbl 1345.62077)]. The authors present a computational framework to solve change-point detection problems which is completely general: it can process many high-dimensional situations achieving improved false detection control. Change-point detection can be viewed as a multiple-testing problem. The proposed optimization framework is computationally efficient. Moreover, with use of this framework, we can control false detection uniformly according to a pre-specified level. Assuming the observation noises are zero mean sub-Gaussian, the authors develop sequential decision rules and demonstrate that these rules are near-optimal in this context.