Refine
Document Type
- Conference Proceeding (2)
- Master's Thesis (1)
- Working Paper (1)
Has Fulltext
- yes (4)
Is part of the Bibliography
- yes (4) (remove)
Keywords
Publicationstate
Reviewstate
We describe a systematic and application-oriented approach to training and evaluating named entity recognition and classification (NERC) systems, the purpose of which is to identify an optimal system and to train an optimal model for named entity tagging DeReKo, a very large general-purpose corpus of contemporary German (Kupietz et al., 2010). DeReKo 's strong dispersion wrt. genre, register and time forces us to base our decision for a specific NERC system on an evaluation performed on a representative sample of DeReKo instead of performance figures that have been reported for the individual NERC systems when evaluated on more uniform and less diverse data. We create and manually annotate such a representative sample as evaluation data for three different NERC systems, for each of which various models are learnt on multiple training data. The proposed sampling method can be viewed as a generally applicable method for sampling evaluation data from an unbalanced target corpus for any sort of natural language processing.
The task-oriented and format-driven development of corpus query systems has led to the creation of numerous corpus query languages (QLs) that vary strongly in expressiveness and syntax. This is a severe impediment for the interoperability of corpus analysis systems, which lack a common protocol. In this paper, we present KoralQuery, a JSON-LD based general corpus query protocol, aiming to be independent of particular QLs, tasks and corpus formats. In addition to describing the system of types and operations that Koral- Query is built on, we exemplify the representation of corpus queries in the serialized format and illustrate use cases in the KorAP project.
The present thesis introduces KoralQuery, a protocol for the generic representation of queries to linguistic corpora. KoralQuery defines a set of types and operations which serve as abstract representations of linguistic entities and configurations. By combining these types and operations in a nested structure, the protocol may express linguistic structures of arbitrary complexity. It achieves a high degree of neutrality with regard to linguistic theory, as it provides flexible structures that allow for the setting of certain parameters to access several complementing and concurrent sources and layers of annotation on the same textual data. JSON-LD is used as a serialisation format for KoralQuery, which allows for the well-defined and normalised exchange of linguistic queries between query engines to promote their interoperability. The automatic translation of queries issued in any of three supported query languages to such KoralQuery serialisations is the second main contribution of this thesis. By employing the introduced translation module, query engines may also work independently of particular query languages, as their backend technology may rely entirely on the abstract KoralQuery representations of the queries. Thus, query engines may provide support for several query languages at once without any additional overhead. The original idea of a general format for the representation of linguistic queries comes from an initiative called Corpus Query Lingua Franca (CQLF), whose theoretic backbone and practical considerations are outlined in the first part of this thesis. This part also includes a brief survey of three typologically different corpus query languages, thus demonstrating their wide variety of features and defining the minimal target space of linguistic types and operations to be covered by KoralQuery.