Issue |
Nat. Sci. Soc.
Volume 29, Number 2, Avril/Juin 2021
|
|
---|---|---|
Page(s) | 223 - 232 | |
Section | Regards – Focus | |
DOI | https://doi.org/10.1051/nss/2021036 | |
Published online | 10 September 2021 |
Comment mobiliser des approches de fouille de textes et d’extraction de la terminologie dans un contexte pluridisciplinaire ?
How to integrate text-mining and terminology extraction approaches in a multidisciplinary context?
Informatique, CIRAD, UMR TETIS (Université de Montpellier, AgroParisTech, CIRAD, CNRS, INRAE),
Montpellier, France
* Auteur correspondant : mathieu.roche@cirad.fr
L’analyse des masses de données nécessite l’utilisation de méthodes mêlant harmonieusement différentes disciplines comme l’informatique, les mathématiques, les statistiques. L’ensemble de ces méthodes utiles pour traiter de telles données forme le socle de la « science des données ». Dans ce cadre, les approches de fouille de textes permettent de découvrir des connaissances utiles et nouvelles pour des experts issus généralement de différents domaines d’application (par exemple, veille épidémiologique, sécurité alimentaire, etc.). Cet article dresse un panorama de l’utilisation de méthodes de fouille de textes dans différents projets liés à l’agriculture et à la santé. Une démarche méthodologique générique est ensuite proposée et discutée.
Abstract
The analysis of large amounts of data requires using methods that combine a range of fields such as computer science, mathematics, statistics, etc. All of these methods useful for data processing form the basis of “data science”. In this context, text mining approaches allow the discovery of new and useful knowledge for experts generally originating from different application areas (e.g., epidemiological surveillance, food security, etc.). This paper provides an overview of the use of text mining methods in various projects related to agriculture and health. A generic methodological approach is then proposed and discussed. This is based on three stages:
- (1)
Data collection. Corpus acquisition from the Web can be done with queries on search engines or RSS feeds.
- (2)
Extraction of terminology using text-mining approaches. Terminology is automatically extracted using different parameters of the BioTex tool (e.g., F-TFIDF-C and C-value measures) dealing with texts in English, French and Spanish.
- (3)
Validation of terms with end-users and field experts based on different approaches (e.g., surveys, workshops, etc.).
Mots clés : fouille de textes / terminologie / corpus / agriculture / santé
Key words: text-mining / terminology / corpus / agriculture / health
© M. Roche, Hosted by EDP Sciences, 2021
This is an Open Access article distributed under the terms of the Creative Commons Attribution License CC-BY (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, except for commercial purposes, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.