Forecasting Emergency Medical Service Call Arrival Rates

• Activity Number: 307
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): David S. Matteson*+ and Mathew W. McLean and Dawn B. Woodard and Shane Henderson
• Companies: Cornell University and Cornell University and Cornell University and Cornell University
• Keywords: Ambulance Planning ; Dynamic Factor Model ; Forecasting ; Inhomogeneous Poisson Process ; Integer-valued Time Series ; Smoothness Constraints

Abstract

We consider a dynamic factor model for predicting the latent rate of ambulances call arrivals which we regard as a time-inhomogeneous Poisson process. Our primary objective is to estimate the regular pattern of temporal variation in emergency priority call arrivals and provide parsimonious forecasts of long and short term ambulance demand. Smoothness constraints are incorporated into estimating both the factor levels and loadings to improve long term forecasting relevant for staff scheduling. Various integer-valued time series models are considered for integrating the remaining residual dependence into adaptive short term forecasts for improved real-time resource management. Our empirical results demonstrate how forecasting ambulance demand can reduce response times and operating costs.