Published June 17, 2019 | Version v1.0.0
Software Open

dice-group/rich-context-competition: Phase-2 Final submission code

  • 1. University of Paderborn
  • 2. Leipzig University
  • 1. University of Paderborn

Description

The repo contains our final submission's refactored code for the Rich Context Competition.

On a high-level, it contains modules for 3 different components of the pipeline:

  1. Preprocessing 
  2. Dataset identification using a combination of simple dataset-search and CRFs based on Rasa-NLU
  3. Identification of methods and fields using vectorization and cosine similarity

Files

dice-group/rich-context-competition-v1.0.0.zip

Files (44.7 MB)

Name Size Download all
md5:ef5c3a8cd2eca27615786b0187113bee
44.7 MB Preview Download

Additional details