A0936
Title: Uniform variance reduced simultaneous inference of time-varying correlation networks
Authors: Lujia Bai - Tsinghua University (China) [presenting]
Weichi Wu - Tsinghua University (China)
Abstract: A flexible framework is proposed for inferring large-scale time-varying and time-lagged correlation networks from non-stationary multivariate or high-dimensional non-stationary time series with piecewise smooth trends. Built on a novel and unified multiple-testing procedure of time-lagged cross-correlation functions with a fixed or diverging number of lags, the method can accurately disclose flexible time-varying network structures associated with complex functional structures at all time points. The applicability of the method is broadened to the structure breaks by developing difference-based nonparametric estimators of cross-correlations, accurate family-wise error control is achieved via a bootstrap-assisted procedure adaptive to the complex temporal dynamics, and the probability of recovering the time-varying network structures is enhanced using a new uniform variance reduction technique for simultaneous inference of nonparametric estimate, which is of independent interest. The asymptotic validity of the proposed method is proved, and its effectiveness is demonstrated in finite samples through simulation studies and empirical applications.
