Node for German Sentiment Analysis


I am trying to perform a sentiment analysis of German texts with the Knime text processing extension. I know that I need to "train" the tag filter with a labelled German language dictionary to be able to conduct the analysis, however I have no idea how to do that...

Can somebody explain how it works?

Thank you!


Please have a look at this example:

In short, you need to provide the Dictionary Tagger with sentiment dictionaries. Note that you will have to use one tagger node per sentiment (e.g. one node for positive terms, one for negative terms).

You can then do the classification as shown here:

Hope that helps!




you can do sentiment analysis in two ways.

  1. By providing a list of negative and positive words. Use these two lists to tag the documents using the Dictionary Tagger node. Then count the positive and negative words and decide on the frequencies of the document is positive or negativ e.g. by a simple Rule Engine node.

  2. By machine learning and predictive analytics. Therefore please see the blog post about sentiment analysis.
    For this approach you need labeled data i.e. documents labeled as positive or negative,

Cheers, Kilian