A text retrieval method for the European Commission's MARS database: selecting human error related accidents

In a world of 'over-information' (much gossip, little knowledge), successful information management increasingly depends on the use of sophisticated methods to effectively extract relevant knowledge from that already acquired on some topic and in some format, for application according to current topics and formats of interests. In this paper, a quite general approach to knowledge elicitation from industrial accidents databases is presented and its implementation in the European Commission's Major Accident Reporting System (MARS) database is discussed. Its practical applicability is demonstrated by deriving lessons learned from major industrial accidents notified to MARS from member states of the European Union under the mandatory requirements of the 'Seveso Directive' with regard to the contribution of human errors to such events. (C) 1999 Elsevier Science Ltd. All rights reserved.