Abstract:
The problem of derived words in automatic text processing based on morphological electronic dictionaries is discussed in this paper. The problem of unknown words (words that are not in e-dictionaries) produced by using derivational patterns is discussed first and the mechanism of processing them which uses morphological grammars as the enhancement of regular expressions is presented. Further, the possibilities of the enhancement of the lemma description are analyzed that would comprise derivatives of a given lemma. Two possible enhancements are suggested, their characteristics are presented as well as directions for implementation.
