Support for interactive document selection in cross-language information retrieval

As digital libraries grow to global scale, the provision of interactive acc ess to content in many languages will become increasingly important. In sys tems that support query-based searching, the presence of multilingual conte nt will affect both the search technology itself and the user interface com ponents that support query formulation, document selection and query refine ment. This article describes the interactions among these components and pr esents a practical way of evaluating the adequacy of the selection interfac e. A categorization-based model for the user's selection process is present ed and an experimental methodology suitable for obtaining process centered results in this context is developed. The methodology is applied to assess the adequacy of a selection interface in which multiple candidate translati ons for a term can be simultaneously presented. The results indicate that t he modeled selection process is somewhat less effective when users are pres ented with multi-translation glosses from Japanese to English rather than m aterials generated originally in English, but that users with access to the gloss translations substantially outperform a Naive Bayes classification a lgorithm. (C) 1999 Elsevier Science Ltd. All rights reserved.