A modular sequence of retrieval procedures to delineate a scientific field: from vocabulary to citations and back

This communication presents a modular arrangement of lexical and citation operations to achieve a satisfactory delineation of a scientific field. Three querying methods are considered: on journals, on articles vocabulary, and on citation network. Rather than associating these querying modes, we consider possible sequences that make the best use of each method. At any stage of iteration, the retrieved set can be enhanced by either a citation-based analysis, or a vocabulary analysis (relative dictionaries), in order to reduce silence or noise. General noisereduction techniques, such as clustering, are applicable at various points of the procedure. A particular sequence on a complex field (nanosciences) is described, starting with journal and lexical query, then applying a citation expansion with a final lexical adjustment. Another sequence is sketched, starting with an additional journalbased procedure.