Modeling the risk of structural fire incidents using a self-organizing map

A Self-Organizing Map (SOM) is used to classify and assess the risk levels of structural fire incidents. Such an assessment can be used not only for improving fire safety and protection of existing and future structures, but also for enhancing emergency responses to future fire incidents. This can reduce the damages and injuries resulting from fires. The map has a 2D hexagonal lattice structure and was applied on a sample of structural fire incident records of Toronto which were reported between 2000 and 2006. Assessment results suggest that the SOM approach is able to successfully classify incidents with different properties into their predefined risk level classes. In summary, the proposed approach shows superior performance for predicting risk, although the quality and quantity of training samples is critical to the success of the predictions