Efthimiadis, E.N. "Data/Document Fusion in Information Retrieval." In: Maxian, B., ed. ASIS'94, Proceedings of the 57th American Society for Information Science (ASIS) Annual Meeting, October 17-20, 1994, Alexandria, VA. New Jersey: Learned Information, vol.31, 1994, p98. [Abstract]
Data Fusion (DF) is a broad discipline encompassing a variety of techniques for combining data, or evidence, from different sources to achieve a unified perspective on some object, or event, of interest (Hall 1992). Although the term "data fusion" or "document fusion" has not often been used in discussions of information retrieval, most research on probabilistic or vector space models, as well as approaches using natural language understanding techniques can be regarded as instances of data fusion. Recently data/document fusion has been explicitly considered in some research which has attempted to fuse the results of several different retrieval methods to produce a final ranking of documents (Belkin et al., 1994, Belkin et al., 1993, Belkin, Kantor & Cool, 1993, Fox and Shaw 1993, Thompson 1993a,b).
The papers in this session describe different approaches to data/document fusion for information retrieval. These approaches combine the output of individual retrieval systems, algorithms, or queries.