Inproceedings,
Using Syntactic Dependency and Language Model X-IOTA IR System for CLIPS Mono and Bilingual Experiments in CLEF 2005
Workshop CLEF 2005, Vienne, Autriche, (2005)
Abstract
This document describes the CLIPS experiments done for the CLEF 2005 campaign. We use surface-syntactic parser in order to extract new indexing terms. These terms are syntactic dependencies. Our goal is to evaluate their interest for an information retrieval task. We used them under different forms in different information retrieval models, particularly in a language model. For the bilingual part we tried two simple tests on Spanish and German to French evaluation, for the translation we use a lemmatization and a dictionary.
