Enriching a Portuguese WordNet using Synonyms from a Monolingual Dictionary

Enriching a Portuguese WordNet using Synonyms from a Monolingual Dictionary Alberto Sim˜oes1,3, Xavier G´omez Guinovart2, Jos´e Jo˜ao Almeida3 1Centro de Estudos Human´ısticos, Universidade do Minho, Portugal 2Universidade de Vigo, Spain 3Centro Algoritmi, Universidade do Minho, Portugal ambs@ilch.uminho.pt, xgg@uvigo.es, jj@di.uminho.pt Abstract In this article we present an exploratory approach to enrich a WordNet-like lexical ontology with the synonyms present in a standard monolingual Portuguese dictionary. The dictionary was converted from PDF into XML and senses were automatically identified and annotated. This allowed us to extract them, independently of definitions, and to create sets of synonyms (synsets). These synsets were then aligned with WordNet synsets, both in the same language (Portuguese) and projecting the Portuguese terms into English, Spanish and Galician. This process allowed both the addition of new term variants to existing synsets, as to create new synsets for Portuguese. Keywords: WordNet, lexical ontology, synonym dictionary, translation, lexical acquisition 1. Introduction There are various initiatives aiming at the construction of a Portuguese lexical ontology, with different goals and some substantial differences in their building pro- cess. Their authors have been working together in order to reuse the other authors’ results to enrich their own re- sources (Gonc¸alo Oliveira et al., 2015). Nevertheless, each one of these initiatives continue working independently. This article presents an experiment being performed by one of these projects: PULO. PULO1 (Sim˜oes and G´omez Guinovart, 2014), that stands for Portuguese Unified Lexical Ontology, aims at integrat- ing linguistic resources of diverse types into a WordNet-like structure aligned with Princeton WordNet (Miller, 1995). Previous work has been done in order to expand PULO (Sim˜oes and Almeida, 2015), but it lacks both qual- ity and coverage. We will discuss two different methodologies to enlarge and, when possible, refine PULO, using synonyms in- formation obtained from a standard monolingual dictio- nary (Casteleiro, 2006). This dictionary include synonyms aligned with each word sense, which means that it is pos- sible to extract, from the same dictionary entry, different synonyms for different senses. Thus, when creating sets of synonyms for each sense we obtain something very similar to the synsets of WordNet. These synsets were then aligned with the synsets already present in PULO, allowing the addition of new variants of the sense2. Moreover, these synsets were also trans- lated into Galician, English and Spanish, and their transla- tions were aligned with the synsets of the wordnets of these languages gathered in the Multilingual Central Repository (MCR) (Gonzalez-Agirre and Rigau, 2013). The fact that PULO is aligned with these other wordnets through the In- terLingual Index (ILI) allows the creation of new synsets and the addition of new variants to existing synsets. 1http://wordnet.pt/ 2We use the term variant to refer to each one of the synonyms in a synset. A small evaluation on the results presents about 80% of pre- cision for variants in new synsets, and about 70% of preci- sion for new variants for existing synsets. This article is structured as follows: Section 2. discusses similar approaches, both for other languages …

Content truncated to 500 words