A0373
Title: Time-varying generalized network autoregressive models
Authors: Boyao Wu - University of International Business and Economics (China) [presenting]
Abstract: A novel class of time-varying network autoregression models is considered to extend popular network autoregressive models by allowing for general network structures, time-varying model coefficients, and the cross-sectionally dependent error term. A local linear method is proposed to estimate time-varying coefficients, and a recursive-design bootstrap procedure is developed to construct valid confidence intervals for time-varying coefficients in the presence of the cross-sectional dependent error term. Asymptotic theories are established on the estimate and the bootstrap procedure under mild conditions. The proposed estimation and bootstrap procedure are illustrated using simulated and real data. The main contribution is to linear models with the network effect, and light is shed on bootstrap inferences and locally stationary processes.
