CMStatistics 2021: Start Registration
View Submission - CMStatistics
B0974
Title: Functional dynamic prediction modeling for hourly ambulance demand data Authors:  Toshihiro Misumi - Yokohama City University (Japan) [presenting]
Abstract: A dynamic prediction problem for the hourly ambulance demand in Yokohama City, Japan, is considered. We propose a novel functional dynamic prediction modeling based on the functional response regression with both considering long-term trends and seasonal variations. The proposed model enables us to dynamically predict the hourly emergency ambulance demands. We apply the proposed method to the analysis of emergency ambulance demand of Yokohama City from 2008 to 2019. The effectiveness of the proposed method is investigated by Monte Carlo simulations. Our proposed method shows a higher prediction result compared to the existing models without long-term trend and seasonality variations.