Intelligent query construction for multilingual document exploitation

Multilingual document exploitation (MDE) involves assessing the relevance of individual foreign language documents to the course of a military mission. A critical task for users of the RAVEN system is the construction of queries that are adequately effective in identifying the foreign language documents most relevant to their information needs. The paper proposes methods to assist the user in: (i) constructing an ontology based query and (ii) determining fuzzy importance weights for query concepts. The proposals draw from information retrieval (IR) research on query expansion, query refinement, query term relevance weighting as well as the sub-field of fuzzy information retrieval. The query construction methods for RAVEN are presented in the context of this IR research.