Awesome

TUWienKBS at GermEval 2018: German Abusive Tweet Detection

GermEval Task 2018 - Shared Task on the Identification of Offensive Language

This repository contains the code for our submission to the GermEval 2018 Shared Task.

Our approach ranked:

• 1st on Task I - Binary classification (out of 51 submissions from 20 teams)
• 2nd on Task II - Fine-grained classification (out of 25 submissions from 11 teams)

A complete description of our system can be found in the Proceedings of the GermEval 2018 Workshop.

Requirements

We used the pretrained word2vec-style model

twitter-de_d100_w5_min10.bin

provided by Josef Ruppenhofer (Heidelberg University).

The file (along with more information about it) can be found here.

Contact persons

Joaquín Padilla Montani, jpadillamontani@gmail.com
Peter Schüller, ps@kr.tuwien.ac.at