Representing the Context of Equivalent Query Words as a Means of Preserving Search Precision
Abstract
Abstract. Information retrieval applications utilizing equivalent terms (e.g. synonyms, alternative spellings. expansions of acronyms and abbreviations), face the problem that the appropriateness of polysemous candidate terms cannot be known a priori. This paper describes a method of utilizing the context of the query words in order to provide a clue to selecting, from among the possible equivalent terms, the ones that have a higher utility, where utility is defined as likeliness to improve recall while preserving precision. A method of “seeding” the system with thousands of potential contextual relations enables a rapid runtime assignment of context to the user’s present query. The result is the broadening of retrieval through equivalent terms, without the typical concomitant decrease in precision. 1 1.
Cite
Text
Musgrove. "Representing the Context of Equivalent Query Words as a Means of Preserving Search Precision." International Joint Conference on Artificial Intelligence, 2005.Markdown
[Musgrove. "Representing the Context of Equivalent Query Words as a Means of Preserving Search Precision." International Joint Conference on Artificial Intelligence, 2005.](https://mlanthology.org/ijcai/2005/musgrove2005ijcai-representing/)BibTeX
@inproceedings{musgrove2005ijcai-representing,
title = {{Representing the Context of Equivalent Query Words as a Means of Preserving Search Precision}},
author = {Musgrove, Timothy A.},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2005},
url = {https://mlanthology.org/ijcai/2005/musgrove2005ijcai-representing/}
}