Publication at Faculty of Mathematics and Physics |
2006
Abstract
Prague Dependency Treebank (PDT) is one of the most important language corpora in the world. The aim of our work is to introduce a software system which builds an exercise book of Czech language upon the PDT data.
Two kinds of exercises are provided, morphological (classifying parts of speech and morphological categories of words) and syntactic (parsing a sentence and classifying syntactic functions of words). Due to the differences between the academic approach and the school approach to the parsing of sentences, the PDT data cannot be used directly.
Many sentences have to be discarded completely, several transformations need to be applied to the others in order to convert them to a form the students are familiar with.