This paper describes the Darmstadt University of Applied Sciences (h da) submission to the binary classification task of the Germeval Task 2018 - Shared Task on the Identification of Offensive Language. We submitted three runs, both a combination of lexical resources and an adapted sentiment analysis system. In run 1 (hda coarse 1.txt) and 2 (hda coarse 2.txt), we had a thresholdbased approach (with different thresholds) and in run 3 (hda coarse 3.txt), an approach based on machine learning.