We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Named Entity Normalization: Combining Normalization Rules, Endogenous Resources and User-Oriented Process.
- Authors
Andréani, Vanessa; Roy, Thibault; Lebarbé, Thomas
- Abstract
Normalization is involved in many fields of information processing. It improves the performance of several applications, such as information retrieval or information extraction, and makes the construction of language resources more reliable. Normalization consists in standardizing each variant of a term or named entity into a unique form, and in this way restricts the impact of language variation. Our work applies to named entity normalization, and aims at optimizing fine-grained corpus analyses carried out by the TecKnowMetrix Company. Our approach mixes several methods, such as pattern matching, similarity metrics and endogenous techniques. Moreover, we place the user in the center of our normalization process, in order to obtain fully reliable data that fit his or her needs.
- Subjects
TECKNOWMETRIX SAS; ASSISTED searching (Information retrieval); USER-generated content; INFORMATION retrieval; INFORMATION resources management; INFORMATION services
- Publication
Canadian Journal of Information & Library Sciences, 2011, Vol 35, Issue 3, p229
- ISSN
1195-096X
- Publication type
Article