POS Tagger and Stanford Tagger

Hi all!

I want to analyze German texts with KNIME. Do I need the POS Tagger first followed by the Stanford Tagger or only the Stanford Tagger for German Language?

Thanks in advance!

Hey @Stol,

welcome to the KNIME forum!

The POS Tagger is using the OpenNLP library and only provides a model for English texts. So for better results, I’d recommend to use the Stanford Tagger with a model for German.

Cheers,

Julian

2 Likes

Thank you for your reply and your hint! In the attached picture, both nodes are used successively. But it makes more sense to use only the Stanford Node for the enrichment.

Hey @Stol,

you can use both nodes successively, but for POS tagging it’s not always necessary. I’d go for the node that performs better on your texts which would be the Stanford tagger anyway, since there are dedicated models for German language available. If you want to tag English texts and you’re unsure which one to choose you can use both ones. The order doesn’t matter since they don’t rely on the output of each other.

Cheers,
Julian

3 Likes