in Computational Linguistics

ii The thesis presents the analysis of dependency parsing systems that are used for parsing German. We restrict ourselves to four parsers operated on the same input data: two of the parsing systems are data-driven; other two are grammar-driven parsers. The main similarity of these parsers is that all of them are based on the same linguistic framework – the dependency theory described in the second chapter of the thesis. In the next chapters we analyze basic features, behavior and performance of each parsing system separately and the considerable part of the thesis is dedicated to the description of errors made by the parsers. Then we summarize their performance, their drawbacks and advantages in comparison to each other. One of our tasks is to show the ways for improving the performance of each parser. Even the MST parser, that shows the best parsing results (about 90 % accuracy), can be improved. For future work we suggest to combine the described parsers as one of the possible ways to achieve better parsing results