For fire departments, having enough firefighters available during a shift is obviously an important requirement. Nevertheless, just like in any organization, having too many firefighters standby is not desirable from a financial point of view. Despite the fact that fire departments can and should not be run like production companies, at least for staffing purposes, forecasting the number of incidents that each fire station has to handle is highly relevant. In this paper, we develop models to create a forecast for the number of incidents that each fire station in the Dutch safety region Amsterdam-Amstelland has to handle for specific incident types and deal with major and small incidents. Previous studies mainly focused on multiplicative models containing correction factors for the weekday and time of year. Our main contribution is to incorporate the influence of different weather conditions in the categories of wind, temperature, rain, and visibility. Rain and wind typically have a strong linear influence, while temperature mainly has a non-linear influence. We show that an ensemble model has the best predictive performance.

, , , ,
International Journal on Advances in Intelligent Systems

Legemaate, G.A.G, de Deijn, J, Bhulai, S, & van der Mei, R.D. (2021). Severe weather-based fire department incident forecasting. International Journal on Advances in Intelligent Systems, 14, 14–23.