Exploiting a Multilingual Web-based Encyclopedia for Bilingual Terminology Extraction

Exploiting a Multilingual Web-based Encyclopedia for Bilingual Terminology Extraction Fatiha Sadat University of Quebec in Montreal, Computer Science department, 201 President Kennedy avenue, Montreal, QC, Canada sadat.fatiha@uqam.ca Abstract. Multilingual linguistic resources are usually constructed from parallel corpora, but since these corpora are available only for selected text domains and language pairs, the potential of other resources is being explored as well. This article seeks to explore and to exploit the idea of using multilingual web-based encyclopedias such as Wikipedia as comparable corpora for bilingual terminology extraction. We propose an approach to extract terms and their translations from different types of Wikipedia link information and data. The next step will be using a linguistic-based information to re-rank and filter the extracted term candidates in the target language. Preliminary evaluations using the combined statistics-based and linguistic-based approaches were applied on different pairs of languages including Japanese, French and English. These evaluations showed a real open improvement and a good quality of the extracted term candidates for building or enriching multilingual ontology, dictionaries or feeding a cross-language information retrieval system with the related expansion terms of the source query. Keywords: Bilingual terminology, comparable corpora, Wikipedia, multilingual linguistic tool. 1 Introduction In recent years two types of multilingual corpora have been an object of studies and research related to natural language processing and information retrieval: parallel corpora and comparable corpora. The parallel corpora are made up of original texts and their translations (Morin et al., 2004 ; Véronis, 2000). This allows texts to be aligned and used in applications such as computer-aided translator training and machine translation systems. This method could be expensive for any pair of languages or even not applicable for some languages, which are characterized by few amounts of Web pages on the Web. On the other hand, non-aligned comparable corpora, more abundant and accessible resources than parallel corpora, have been given a special interest in bilingual terminology acquisition and lexical resources enrichment (Dejean et al., 2002; Fung, 2000; Gœuriot et al., 2009a; Gœuriot et al., 2009b; Morin et al., 2006; Nakagawa et al., 2000; Rapp, 1999; Sadat et al., 2003; Sadat et al., 2004). Comparable corpora are defined as collections of texts from pairs or multiples of languages, which can be contrasted because of their common features in the topic, the domain, the authors, the time period, etc. Comparable corpora could be collected from downloading electronic copies of newspapers and articles, on the WWW for any specified domain. Among the advantages of comparable corpora; their availability, consistency and utility for research on Natural Language Processing (NLP). In another hand, recent publications on bilingual terminology extraction from comparable corpora have shown promising results although most used comparable corpora are domain-specific, which causes limitations on the usage diversity, the domain and the quality of terminology. This paper intends to bring solutions to the problem of lexical coverage of existing linguistic resources such as multilingual ontologies and dictionaries, but also to the improvement of the performance of Cross-Language Information Retrieval. The main contribution …

Content truncated to 500 words