Extracting recurrent scenarios from narrative texts using a Bayesiannetwork: Application to serious occupational accidents with movement disturbance

A probabilistic approach has been developed to extract recurrent serious Occupational Accidents with Movement Disturbance (OAMD) scenarios from narrative texts within a prevention framework. Relevant information in 143 texts was first coded by logical combinations of generic accident factors. A Bayesian network (BN)-based model was then built for OAMDs using these data and the expert knowledge. A data clustering step was subsequently performed to group the OAMDs into similar classes from a generic factor occurrence and a pattern standpoint. Finally, the Most Probable Explanation (MPE) was derived as the associated recurrent scenario for each class. Eight scenarios were thus extracted to describe the 143 OAMDs in the construction and metallurgy sectors. Their recurrent nature is discussed.
Probable combinations of generic factors provide a fair representation of particularly serious OAMDs, as described in narrative texts. This is a useful contribution to making companies aware of the variety of circumstances, in which these accidents occur, to progressing in the prevention of such accidents and to developing an analysis framework dedicated to this kind of accident.