CLARIN and Free Open Source Finite-State Tools

Abstract

A new emerging European research infrastructure called CLARIN and a related project called HFST are briefly described. HFST has built a programming interface on top of some existing open source finite-state packages such as SFST and OpenFST. In order to verify its utility, HFST has built open source tools on top of this HFST interface. These tools create lexical transducers, compile morphophonological two-level rules and combine them into a transducer lexicon. The tools have been tested against independently created with full-scale lexicons and rules for Northern Sami and Lule Sami languages which have more complicated lexical and morphophonological structure than most other European languages.