A knowledge-based fatal incident decision model

摘要

A methodology for determining remains identification (ID) following a mass disaster is presented. The solution methodology is domain-independent and capable of addressing a wide range of assignment problems. A knowledge-based fatal incident decision model (FINDM) for providing a decision support to forensic scientists involved in the skeletal ID process is discussed. A mathematical framework for FINDM is developed that integrates a knowledge base with a network flow algorithm for resolving conflicts during the ID process. The FINDM framework has been implemented can an IBM PC and includes an observation advisor, an assignment advisor, and a conflict resolution module. Knowledge acquisition and representation issues are discussed, along with a numerical example and results. With respect to the remains ID problem, the FINDM approach shifts major efforts in resolving the problem from that of establishing a method of assignment to that of controlling the quality of data collected, improving domain knowledge, and analyzing conflicts.