Semantic Role Labeller

Semantic Role Labelling (SRL) for Dutch and English documents.

Inputs

  • Corpus: A dataset of one or more textual documents in Dutch or English.

Outputs

  • Table: A data table providing all frame information (subject, verb, object) tuples after performing SRL analysis. We use the implementation by Jisk Attema and Dafne van Kuppevelt at https://github.com/Filter-Bubble/stroll

SRL Parser widget applies SRL to decompose sentences into (subject, verb, object) components of narrative. It outputs a table storing all the tuple information.

../_images/dsgsrl-example.png

Example

This simple example shows how SRL Parser can enrich the workflow. We have used a dataset of Dutch narrative stories about the corona pandemic.