Smoothing of Correlation Functions in the Estimation of the Order of Influenza Infection between Adults and Children

Abstract

The cross-correlation functions between the time series of drug sales at community pharmacies have been used for the epidemiological understanding of influenza propagation process. The aim of this paper is to examine the methods of smoothing for determining the maximum position (time lag) of the cross-correlation function. The moving average method with a window of seven days is taken to eliminate the hebdomadal frequencies originating from the life style of people around the pharmacies, or rather, purchasers. Based on the smoothed data of influenza anti-viral agents for adults and children at pharmacies in Tokyo and its vicinity, this paper tries to give an answer to the question of which are earlier infected with influenza, adults or children, in the limited areas.