German Hate Speech Detection on Twitter

This paper describes our system submission for the GermEval 2018 shared task on the identification of German hate speech in Tweets at Konvens 2018. We trained and tested a Logistic Regression classifier with 10-fold cross validation using character ngrams as features. We achieved a macro F1 of 76.72 for the coarse-grained classification task and 47.17 for the fine-grained task when testing the classifiers on a small development set we created