Cluster analysis for road accidents investigation

Road accidents and particularly urban accidents are one of the most important negative impacts produced by travel, involving both users and non users of the transport system. The number of fatal injuries has increased in the last decades and their social cost has become more and more relevant requiring an in-depth study in order to resolve the problem, This paper proposes an analysis of accidents based on cluster techniques. Cluster techniques need a suitable data-base in order to group sets of “similar” objects and to identify the most relevant elements that represent the group. The identification of the relevant aspects common to different types of accidents is the first step for a conscious intervention in the transportation system, because the knowledge of the principal causes of accidents can help the analysts of the transportation systems both in the construction of mathematical relationships among accident and causes and can support the choice of suitable actions for reducing the number of accidents (mainly the fatal accidents).