Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
This paper reports a pilot study, in which Constraint G r a m m a r inspired rules were learnt using the Progol machine-learning system. Rules discarding faulty readings of ambiguously tagged words were learnt for the part of speech tags of the Stockholm-Ume£ Corpus. Several thousand disambiguation rules were induced. When tested on unseen data, 98% of the words retained the correct reading after tagging. However, there were ambiguities pending after tagging, on an average 1.13 tags per word. | Learning Constraint Grammar-style disambiguation rules using Inductive Logic Programming Nikolaj Lindberg Centre for Speech Technology Royal Institute of Technology SE-100 44 Stockholm Sweden nikolaj@speech.kth.se Abstract This paper reports a pilot study in which Constraint Grammar inspired rules were learnt using the Progol machine-learning system. Rules discarding faulty readings of ambiguously tagged words were learnt for the part of speech tags of the Stockholm-Umea Corpus. Several thousand disambiguation rules were induced. When tested on unseen data 98 of the words retained the correct reading after tagging. However there were ambiguities pending after tagging on an average 1.13 tags per word. The results suggest that the Progol system can be useful for learning tagging rules of good quality. 1 Introduction The success of the Constraint Grammar CG Karlsson et al. 1995 approach to part of speech tagging and surface syntactic dependency parsing is due to the minutely handcrafted grammar and two-level morphology lexicon developed over several years. In the study reported here the Progol machine-learning system was used to induce CG-style tag eliminating rules from a one million word part of speech tagged corpus of Swedish. Some 7 000 rules were induced. When tested on unseen data 98 of the words retained the correct tag. There were still ambiguities left in the output on an average 1.13 readings per word. In the following sections the CG framework and the Progol machine learning system will be presented very briefly. 1.1 Constraint Grammar POS tagging Constraint Grammar is a system for part of speech tagging and shallow syntactic dependency analysis of unrestricted text. In the fol Martin Eineborg Telia Research AB Spoken Language Processing SE-136 80 Haninge Sweden Martin.E.EineborgQtelia.se lowing only the part of speech tagging step will be discussed. The following as a typical reductionistic example of a CG rule which discards a verbal reading of a word .
