Dieser Webauftritt wurde archiviert und wird nicht mehr aktualisiert.

Bei technischen Fragen kontaktieren Sie bitte das Webteam. Es gilt die Datenschutzerklärung der TU Darmstadt.

Natural Language Processing and Rule-based Information Extraction with UIMA


The tutorial is split into two theoretic parts, followed by a hands-on session.

The first part gives an introduction to doing UIMA-based natural language processing using the DKPro Core component collection.  It will be shown how a processing pipeline can be composed from existing components, such as a tokenizer, part-of-speech tagger, and parser. To illustrate how such a pipeline can be used on own data and how analysis results can be exported to other tools, we also show how to implement simple reader and writer components. This includes a basic introduction to the underlying concepts, such as Maven and uimaFIT.

The second part gives an introduction to Apache UIMA Ruta. We will cover the syntax and semantics of the rule language as well as tooling support for developing rule-based information extraction applications.

In the hands-on session, we will implement and run a simple pipeline with DKPro Core.  We then will extend the pipeline with rule-based post-processing in order to approach and solve different tasks such as information extraction. This includes solutions how to combine the rules with the information annotated by the DKPro components and ways for efficient and effective engineering of the rules themselves.

The tutorial slides and examples are available on the 3rd UIMA@GSCL Workshop website.


Peter Klügl, Dr. Katrin Tomanek, Richard Eckart de Castilho

A A A | Drucken Print | Impressum Impressum | Sitemap Sitemap | Kontakt Contact
zum Seitenanfangzum Seitenanfang