@inproceedings{bansal-etal-2019-tubingen,
    title = {{HAD}-{T}{\"u}bingen at {S}em{E}val-2019 Task 6: Deep Learning Analysis of Offensive Language on {T}witter: Identification and Categorization},
    author = "Bansal, Himanshu  and
      Nagel, Daniel  and
      Soloveva, Anita",
    editor = "May, Jonathan  and
      Shutova, Ekaterina  and
      Herbelot, Aurelie  and
      Zhu, Xiaodan  and
      Apidianaki, Marianna  and
      Mohammad, Saif M.",
    booktitle = "Proceedings of the 13th International Workshop on Semantic Evaluation",
    month = jun,
    year = "2019",
    address = "Minneapolis, Minnesota, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/add-emnlp-2024-awards/S19-2111/",
    doi = "10.18653/v1/S19-2111",
    pages = "622--627",
    abstract = {This paper describes the submissions of our team, HAD-T{\"u}bingen, for the SemEval 2019 - Task 6: {\textquotedblleft}OffensEval: Identifying and Categorizing Offensive Language in Social Media{\textquotedblright}. We participated in all the three sub-tasks: Sub-task A - {\textquotedblleft}Offensive language identification{\textquotedblright}, sub-task B - {\textquotedblleft}Automatic categorization of offense types{\textquotedblright} and sub-task C - {\textquotedblleft}Offense target identification{\textquotedblright}. As a baseline model we used a Long short-term memory recurrent neural network (LSTM) to identify and categorize offensive tweets. For all the tasks we experimented with external databases in a postprocessing step to enhance the results made by our model. The best macro-average F1 scores obtained for the sub-tasks A, B and C are 0.73, 0.52, and 0.37, respectively.}
}