This paper presents an innovative solution for the vulnerability assessment to wildfires, making use of an all-relevant feature selection algorithm established on statistical relationships to develop a physical vulnerability index for buildings subject to wildfire.To prepare for upcoming extreme events, it is critical for decision-makers to have a thorough understanding of the vulnerability of the built environment to wildfire. Building quality and design standards are important not only because building loss is costly but also because robust buildings may offer shelter when evacuation is not possible. However, studies aiming at the analysis of wildfire vulnerability for the built environment are limited.
Data from a recent and systematically documented wildfire event in Greece (Mati, 2018) are used to select and weight the relevant indicators using a permutation-based automated feature selection based on random forests. Building characteristics including the structural type, the roof type, material and shape, the inclination of the ground, the surrounding vegetation, the material of the shutters and the ground covering were selected and formed into the index. The index may be used in other places in Europe and beyond, especially where no empirical data are available supporting decision-making and risk reduction of an emerging hazard amplified by climate change.