JCoRe Part-of-Speech Tagging Analysis Engine
Tagger for annotating a text with (arbitrarily chosen) part-of-speech tags
The JULIE Lab Part of Speech Tagger (JPOS) is a generic and configurable POS tagger. JPOS was tested on the general-language news paper domain and in the biomedical domain; it performs very good for German texts, yet only mediocre for English [HMFH15]. As JPOS employs a machine learning (ML) approach, a model (for the specific domain and entity classes to be predicted) needs to be trained first. Thus, JPOS offers a training mode. Furthermore, JPOS also provides an evaluation mode to assess the current model performance in terms of accuracy.
Requirement and Dependencies
The input and output of an AE is done via annotation objects. The classes corresponding to these objects are part of the JCoRe Type System.
Using the AE - Descriptor Configuration
In UIMA, each component is configured by a descriptor in XML. Such a preconfigured descriptor is available under
src/main/resources/de/julielab/jcore/ but it can be further edited if so desired; see UIMA SDK User's Guide for further information.
2. Predefined Settings
An extensive documentation can be found under
|[HMFH15]||Johannes Hellrich, Franz Matthies, Erik Faessler & Udo Hahn: Sharing Models and Tools for Processing German Clinical Text. In: Ronald Cornet, Lăcrămioara Stoicu-Tivadar, Alexander Hörbst, Carlos Luis Parra Calderón, Stig Kjær Andersen, Mira Hercigonja-Szekeres (Eds.): Digital Healthcare Empowering Europeans [= Proceedings of MIE2015], 2015, pp. 734-738 (Studies in Health Technology and Informatics, 210).|