Authors:
Désiré Kompaoré
1
;
Josiane Mothe
1
and
Ludovic Tanguy
2
Affiliations:
1
Institut de Recherche en Informatique de Toulouse, Université de Toulouse, France
;
2
ERSS, Université de Toulouse, France
Keyword(s):
Information retrieval, data fusion, indexing, information retrieval in French.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Biomedical Engineering
;
Business Analytics
;
Data Engineering
;
Data Mining
;
Databases and Information Systems Integration
;
Datamining
;
Enterprise Information Systems
;
Health Information Systems
;
Information Systems Analysis and Specification
;
Knowledge Management
;
Natural Language Interfaces to Intelligent Systems
;
Ontologies and the Semantic Web
;
Sensor Networks
;
Signal Processing
;
Society, e-Business and e-Government
;
Soft Computing
;
Web Information Systems and Technologies
Abstract:
This paper analyses three type of different indexing methods applied on French test collections (CLEF from 2000 to 2005): lemmas, truncated terms and single words. The same search engine and the same characteristics are used independently to the indexing method to avoid variability in the analysis. When evaluated on French CLEF collections, indexing by lemmas is the best method compared to single words and truncated term methods. We also analyse the impact of combining indexing methods by using the CombMNZ function. As CLEF topics are composed of different parts, we also examine the influence of these topic parts by comparing the results when topic parts are considered individually, and when they are combined. Finally, we combine both indexing methods and query parts. We show that MAP can be improved up to 8% compared to the best individual methods.