This paper describes the work achieved in CESART (Campagne d’Evaluation des Systemes d’Acquisition des Ressources Terminologiques) evaluation project supported by the French Ministry of Research and Technology1 and coordinated by the University of Lille 3 and ELDA. The project deals with the evaluation of term and semantic relation extraction from corpora in French. CESART logically follows on the evaluation project achieved within the framework of the Concerted Research Project ARC A32 supported by the AUF, former Aupelf-Uref. This paper sets the context, briefly mentions the project objectives, reports on the adopted evaluation protocol, describes the evaluation tasks and finally provides the results of the official evaluation campaign. 1 http://www.technolangue.net 2 The ARC A3 is a project of the ILEC group coordinated and founded by AUF 1996-2000. The project aim was to test software capabilities in term and semantic relation extraction from corpora in French (cf. Mustafa El Hadi et al., 1998; 2001).
[1]
Ibekwe-Sanjuan Fidelia,et al.
Mining textual data through term variant clustering: the TermWatch system
,
2004
.
[2]
Widad Mustafa El Hadi,et al.
The ARC A3 Project: Terminology Acquisition Tools: Evaluation Method and Task
,
2001
.
[3]
Christophe Jouis,et al.
Terminology extraction and acquisition from textual data: criteria for evaluating tools and methods
,
1998
.
[4]
Trevor Pinch,et al.
How users matter : The co-construction of users and technologies
,
2003
.
[5]
Eric SanJuan,et al.
Mining Textual Data through Term Variant Clustering : the TermWatch system
,
2004,
RIAO.