Merging thesauri: principles and evaluation

An investigation is reported of ways to take advantage of the semantics in thesauri to improve: (1) indexing by providing descriptions of documents as sets of terms from the thesaurus; and (2) retrieval by assessing the relevance of documents to a query. Thesauri need to be updated to account for the evolution of the field they cover. Accordingly, various augmentation algorithms and methods to assess the usefulness of these augmentations are being studied. The augmentations consist of merging two existing thesauri. By keeping a consistent level of complexity among the structure manipulated by the merging algorithm, the reasoning method, and the evaluation procedure, an improvement of the performance of the merged thesaurus on both document indexing and retrieval is demonstrated.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>