GermaNER

We present GermaNER, the first free (as in: entirely free) and open (as in: open source) named entity recognition tool for German.

While there are plenty named entity recognition tools around, the use of German NER technology was limited to research purposes.

German annotated data for NER, as well as German data and tools for preprocessing, are not obtained easily from the creators, often involving the sending of hardcopy license agreements. Further, despite the projects that created these data and tools have been funded from public sources, it is often not possible to use data and tools for commercial purposes, hampering the development of German language technology as a whole.

We believe that this situation is unfortunate, and we are in the position to change it. This is why we present an open-source, leniently-licensed sequence tagging tool with a pre-trained model for German named entity recognition. The tool neither relies on copyrighted training data nor on preprocessing steps that use restricted software or data.

 

Download and more info on GitHub: https://github.com/tudarmstadt-lt/GermaNER