An inductive reasoning approach for building system safety risk models of aviation accidents

An inductive reasoning approach is employed to develop a prototype hybrid decision support tool whose main objective is to build probabilistic causal models representing the safety risk involved in aviation accidents. In this context, 15 aircraft accidents representative of five major accident types are selected to build an initial seed for the case‐base of the prototype tool. Consequently, within each individual accident model, main clusters of causal factors are identified for inclusion in the initial seed, thereby improving, both quantitatively and qualitatively, the case‐base of the prototype tool. A new methodology developed specifically for indexing aviation accidents into databases is used for indexing the initial seed into the case‐base of the tool. The resulting product is a highly customized conversational decision support tool that provides solution possibilities in the form of probabilistic causal models of accident scenarios retrieved and ranked according to their similarity to the current accident that the intended user investigates.