Optimal Change-point Testing for High-dimensional Linear Models with Temporal Dependence
This paper studies change-point testing for high-dimensional linear models, an important problem that is not well explored in the literature. Specifically, we propose a quadratic-form-based cumulative sum (CUSUM) test to inspect the stability of the regression coefficients in a high-dimensional linear model. The proposed test is able to control the type-I error at any desired level and is theoretically sound for temporally dependent observations. We establish the asymptotic distribution of the proposed test under both the null and alternative hypotheses. Furthermore, we show that our test is asymptotically powerful against multiple-change-point alternative hypotheses and achieves the optimal detection boundary for a wide class of high-dimensional linear models. Extensive numerical experiments and a real data application in macroeconomics are conducted to demonstrate the promising performance and practical utility of the proposed test. In addition, some new consistency results of the Lasso <cit.> are established under the change-point setting with the presence of temporal dependence, which may be of independent interest.
READ FULL TEXT