HOME   /   Content


Zudi Lu: On a Data-driven Semiparametric Model with Penalized Spatio-temporal Lag Interactions
time: Dec 26, 2019

Time:10:00, December 24, 2019 (Tuesday)

Location:1710, building B,Feicuihu Campusscience and education building,FeicuihuCampus

Speaker:ProfessorZudi Lu

From:University ofSouthampton

Organizer:School of Mathematics

Lecturer introduce:Zudi Lu, who received his doctor's degree from the Chinese Academy of Sciences in 1996, is now a chair professor and doctoral supervisor of statistics in the school of Mathematical Sciences and the Southampton Institute of statistical sciences. Research fields include financial statistics, econometrics, nonlinear time series analysis, nonlinear spatiotemporal big data and intelligent modeling. Professor Lu zudi has successively worked in the Institute of mathematics and systems science of the Chinese Academy of Sciences, Catholic University of Leuven in Belgium, London School of economics in the UK, Curtin University in Australia and Adelaide University. He has successively obtained support from China national key natural science fund, arc future fellow of Australian National Research Council and Marie Curie fellow of European Union. He is an elected member of International Statistical Society. More than 80 academic papers have been published in major international statistical and econometric journals, including top journals such as annals of statistics, Journal of American Statistical Association, journal al of Royal Statistical Society Series B, Journal of economics, economic theory, etc. Served as deputy editor of Journal of time series analysis and editorial board member of international journals such as environmental modeling and assessment and cogent mathematics and statistics.

Description:In many applications related to spatial problems, to study possibly nonlinear relationship between covariates and the concerned response at a location, accounting for the temporal lag interactions of the response at a givenlocation and the spatiotemporal lag interactions between locations could improve the accuracy of estimation and forecasting. There lacks, however, methodology to objectively identify and estimate such spatiotemporal lag interactions. In this talk, we present a semiparametric data‐driven nonlinear time series regression method that accounts for l-ag interactions across space and over time. A penalized procedure utilizing adaptive Lasso is dev-eloped for the identification and estimation of important spatiotemporal lag interactions. Theoretical properties for our proposed methodology are established under a general near epoch dependence structure and thus the results can be applied to a variety of linear and nonlinear time series processes. For illustration, we analyze the US housing price data and demonstrate substantial improvement in forecasting via the identification of nonlinear relationship between HPI and CPI as well as spatiotemporal lag interactions.