ABSTRACTS (Chosen by G~ Salton from current issues of journals in the retrieval area) io Experiments in Local Metrical Feedback in Full-Text Retrieva!Systems Rony Attar and Aviezri S. Fraenkel, Department of Applied Mathematics, The Weizmann Institute of Science, Rehovot, Israel A method of iterative searching, using the results of one iteration search to formulate the next iteration search, was applied to a full-text database consisting of some 2400 documents and 1,300,000 text-words of Hebrew and Aramaic. The iterative method consists of clustering the documents returned in an iteration, using weighting by proximity and by frequency simultaneously. This process produces searchonyms, which are terms synonymous to keywords in the context of a single query. Augumenting or replacing keywords by searchonyms via manual or automatic feedback leads to the formulation of the next iteration search. The results of the experiment are consistent with those of an earlier small-scale experiment on an English database, and indicate that in contrast to global clustering where the size of matrices limits applications to small databases and improvements are doubtful, local metrical methods appear to be well suited to arbitrarily large databases, improving precision and recall simultaneously. Further experiments using more test-queries run on even
/lp/association-for-computing-machinery/abstracts-chosen-by-g-salton-from-current-issues-of-journals-in-the-vBmn41PClS