On Relevance, Probabilistic Indexing and Information Retrieval
Journal of the ACM (JACM)
Precision Weighting—An Effective Automatic Indexing Method
Journal of the ACM (JACM)
Information Retrieval Research
Information Retrieval Research
Enhanced hypertext categorization using hyperlinks
SIGMOD '98 Proceedings of the 1998 ACM SIGMOD international conference on Management of data
Global ranking via data fusion
COLING '10 Proceedings of the 23rd International Conference on Computational Linguistics: Posters
Hi-index | 0.00 |
The use of cited title terms of a scientific document for automatic indexing is explored. It offers a means of index term selection as well as term relevance weighting, based on author-provided relevance information and Bayes Theorem as in probabilistic retrieval. The latter quantitative consideration leads to a new measure of document-document similarity measure which is shown to have importance both for initial search and in relevance feedback retrieval, by offering a choice of iterative strategies.Extension of the concept of cited title terms to citing title terms shows that these two approaches are compatible with the current two competing models of probability of relevance for document retrieval (Robertson et al. 1982), if a document can also be regarded as a query. Their term usage may therefore provide the necessary statistics for parameter estimation to test both theories.