HFST—A System for Creating NLP Tools
HFST—a System for Creating NLP Tools Krister Lindén, Erik Axelson, Senka Drobac, Sam Hardwick, Juha Kuokkala, Jyrki Niemi, Tommi A Pirinen, and Miikka Silfverberg University of Helsinki Department of Modern Languages Unioninkatu 40 A FI-00014 University of Helsinki, Finland {krister.linden, erik.axelson, senka.drobac, sam.hardwick, juha.kuokkala, jyrki.niemi, tommi.pirinen, miikka.silfverberg}@helsinki.fi Abstract. The paper presents and evaluates various NLP tools that have been created using the open source library HFST–Helsinki Finite- State Technology and outlines the minimal extensions that this has re- quired to a pure finite-state system. In particular, the paper describes an implementation and application of Pmatch presented by Karttunen at SFCM 2011. Keywords: finite-state technology, language identification, morpholog- ical guessers, spell-checking, named-entity recognition, language genera- tion, parsing, HFST, XFST, Pmatch Introduction In natural language processing, finite-state string transducer methods have been found useful for solving a number of practical problems ranging from language identification via morphological processing and generation to part-of-speech tag- ging and named-entity recognition, as long as the problems lend themselves to a formulation based on matching and transforming local context. In this paper, we present and evaluate various tools that have been created using HFST–Helsinki Finite-State Technology1 and outline the minimal exten- sions that this has required to a pure FST system. In particular, we describe an implementation of Pmatch presented by Karttunen at SFCM 2011 [7] and its application to a large-scale named-entity recognizer for Swedish. The paper is structured as follows: Section 1 is on applications and their evaluation.
[Show full text]