Evaluation of an automatic process for specialized web corpora collection and term extraction for Basque

In this paper we describe the processes for collecting Basque specialized corpora in different domains from the Internet and subsequently extracting terminology out of them, using automatic tools in both cases. We evaluate the results of corpus compiling and term extraction by making use of a specialized dictionary recently updated by experts. We also compare the results of the automatically collected web corpus with those of a traditionally collected corpus, in order to analyze the usefulness of the Internet as a reliable source of information for terminology tasks.