A0602
Title: Convolutional regression for big spatial data
Authors: Yasumasa Matsuda - Tohoku University (Japan) [presenting]
Abstract: It is now common to collect big spatial data on a national or continental scale at discrete time points. The aim is to present a regression model where both dependent and independent variables are big spatial data. Regarding spatial data as functions over a region, we propose a functional regression by a parametric convolution kernel together with the least-squares estimation on the frequency domain by applying Fourier transform, which makes it possible to handle massive datasets with asymptotic validations under the mixed asymptotics. The regression is applied to new weekly cases of coronavirus disease 2019 (COVID-19) and human mobility collected in Japanese cities. We find that an increase in human mobility results in an increase of COVID-19 cases in a time lag of two weeks.
