TBXTools: A Free, Fast and Flexible Tool for Automatic Terminology Extraction

Proceedings of Recent Advances in Natural Language Processing, pages 473–479, Hissar, Bulgaria, Sep 7–9 2015. TBXTools: A Free, Fast and Flexible Tool for Automatic Terminology Extraction Antoni Oliver Universitat Oberta de Catalunya aoliverg@uoc.edu Merc`e V`azquez Universitat Oberta de Catalunya mvazquezga@uoc.edu Abstract The manual identification of terminology from specialized corpora is a complex task that needs to be addressed by flexible tools, in order to facilitate the construction of multilingual terminologies which are the main resources for computer-assisted translation tools, machine translation or ontologies. The automatic terminology extraction tools developed so far either use a proprietary code or an open source code, that is limited to certain software func- tionalities. To automatically extract terms from specialized corpora for different pur- poses such as constructing dictionaries, thesauruses or translation memories, we need open source tools to easily integrate new functionalities to improve term selec- tion. This paper presents TBXTools, a free automatic terminology extraction tool that implements linguistic and statistical methods for multiword term extraction. The tool allows the users to easily iden- tify multiword terms from specialized cor- pora and also, if needed, translation candi- dates from parallel corpora. In this paper we present the main features of TBXTools along with evaluation results for term ex- traction, both using statistical and linguis- tic methodology, for several corpora. 1 Introduction Automatic terminology extraction (ATE) is a rel- evant natural language processing task involv- ing terminology which has been used to iden- tify domain-relevant terms applying computa- tional methods (Oliver et al., 2007a; Foo, 2012). Automatic term extraction is a relevant task that can be useful for a wide range of tasks, such as ontology learning, machine translation, computer- assisted translation, thesaurus construction, classi- fication, indexing, information retrieval, and also text mining and text summarisation (Heid and Mc- Naught, 1991; Frantzi and Ananiadou, 1996; Vu et al., 2008). The automatic terminology extraction tools de- veloped in recent years allow easier manual term extraction from a specialized corpus, which is a long, tedious and repetitive task that has the risk of being unsystematic and subjective, very costly in economic terms and limited by the current available information. However, existing tools should be improved in order to get more consistent terminology and greater productivity (Gornostay, 2010). In the last few years, several term extraction tools have been developed, but most of them are language-dependent: French and English –Fastr (Jacquemin, 1999) and Acabit (Daille, 2003); Portuguese –Extracterm (Costa et al., 2004) and ExATOlp (Lopes et al., 2009); Spanish-Basque –Elexbi (Hernaiz et al., 2006); Spanish-German –Autoterm (Haller, 2008); Arabic (Boulaknadel et al., 2008); Slovene and English –Luiz (Vin- tar, 2010); English and Italian –KX (Pianta and Tonelli, 2010); or English and German (Gojun et al., 2012). Some tools are adapted to a specialized domain: TermExtractor (Sclano and Velardi, 2007), Ter- Mine (Ananiadou et al., 2009) or BioYaTeA (Go- lik et al., 2013), for example. Specific tools have been developed to extract corpus-specific lexical items comparing technical and non-technical cor- pus: TermoStat (Drouin, 2003). And other tools are …

Content truncated to 500 words