Multilingual Terminology Extraction and Validation

Abstract

This paper presents the automatic terminology extraction approach developed within project LIQUID1. This project aims at developing a cost-effective solution for the problem of cross-language access to multilingual text databases in technical and scientific domains. Cross-Language Information Retrieval faces a major challenge: organizing unstructured textual information according to its contents and regardless of its language. Our solution is based on two main components, a terminology extraction tool and a domain-specific ontology. The terminology extraction tool identifies the terminology that describes the contents of a particular document. Then, these terms are linked to a domain-specific ontology. This paper presents the terminology extraction tool and the experimental results obtained in the domain of Gastroenterology.

Resources

Details