Summary: Application of Different Learning Methods to
Hungarian Part­of­speech Tagging
Tam'as Horv'ath 1 , Zolt'an Alexin 2 , Tibor Gyim'othy 3 , and Stefan Wrobel 4;1
1 German National Research Center for Information Technology, GMD ­ AiS.KD,
Schloß Birlinghoven, D­53754 Sankt Augustin, tamas.horvath@gmd.de
2 Dept. of Applied Informatics, J'ozsef Attila University,
P.O.Box 652, H­6701 Szeged, alexin@inf.u­szeged.hu
3 Research Group on Artificial Intelligence, Hungarian Academy of Sciences,
Aradi v'ertanuk tere 1, H­6720 Szeged, gyimi@inf.u­szeged.hu
4 Otto­von­Guericke­Universit¨at Magdeburg, IWS,
P.O.Box 4120, D­39106 Magdeburg, wrobel@iws.cs.uni­magdeburg.de
Abstract. From the point of view of computational linguistics, Hungar­
ian is a difficult language due to its complex grammar and rich morphol­
ogy. This means that even a common task such as part­of­speech tagging
presents a new challenge for learning when looked at for the Hungarian
language, especially given the fact that this language has fairly free word
order. In this paper we therefore present a case study designed to illus­
trate the potential and limits of current ILP and non­ILP algorithms
on the Hungarian POS­tagging task. We have selected the popular C4.5
and Progol systems as propositional and ILP representatives, adding ex­

Source: Alexin, Zoltán - Institute of Informatics, University of Szeged

Collections: Computer Technologies and Information Sciences