Aim: To describe an approach in the classification of critical incidents, which are aggregated for a spatial analysis of threats
Introduction: Classification of critical incidents allows for a capability evaluation of an organized firefighting and rescue system, to meet the demands identified in a given area. It should be based on an analysis of historical data and only such information provides a realistic representation of critical incidents, which occur in a given area. In the context of process stability, this profile can be illustrated by the Poisson distribution and will facilitate short term forecasting. Additionally, such data allows for performance evaluation of the firefighting and rescue system in context of tasks placed upon the system. The critical incident classification, presented in this article, is based on the naive Bayesian codifier, which assumes the independence of classification characteristics. Despite the misleading name and simplified assumptions, it is a very effective tool in the classification process. The article presents methodological assumptions for grading of critical events, describes the algorithm used for the codification of such incidents in the course of spatial threat analysis and provides an example of such an analysis for one of the districts. The study used data stored in the support system for decision making in Command and Control Centres, Fire Bird version 2.5 SWD-ST.
Methodology: Statistical analysis and forecasting.
Conclusions: The distinguishing features of critical incidents are: type of event, operation duration and volume of resources used during incidents. The initial distinction of classification features is linked to the essence of the analysis, whose main objective is to assess the adequacy of potential and timely response of the firefighting and rescue system. Therefore, dominant features for a given category of incident will be: type of event, defined in terms of required equipment and training; the size of incident, defined in terms of the volume of resources necessary to conduct an effective rescue operation. Since the analysis is based on historical data, it is necessary to take into account the volume of resources deployed to a given incident and finally; the duration for the operation, which together with resources used, defines the magnitude of the incident. The analysis and outcomes, of which a limited range is presented in this article, confirms the validity of assumptions.
Keywords: analysis of hazards, statistical analysis, statistical forecasting, statistical classification
Type of article: original scientific article