Mitigating Bias in Session-based Cyberbullying Detection

Implementation of the ACL21 paper: Mitigating Bias in Session-based Cyberbullying Detection

Code usage

The source code is written in Python and is a Jupyter notebook. To use the code:

[1] Install the requirements using the following command:

pip install -r requirements.txt

[2] Please refer to https://jupyterlab.readthedocs.io/en/stable/getting_started/installation.html for installing JupyterLab. After installation, you can use the following terminal command to run JupyterLab and open the shared notebook:

jupyter lab

[3] For the dataset, please refer to the following papers:

[1] Homa Hosseinmardi, Sabrina Arredondo Mattson, Rahat Ibn Rafiq, Richard Han, Qin Lv, and Shivakant Mishra. 2015. Analyzing labeled cyberbullying incidents on the instagram social network. In Socinfo. Springer, 49–66.

[2] Rahat Ibn Rafiq, Homa Hosseinmardi, Richard Han, Qin Lv, Shivakant Mishra, and Sabrina Arredondo Mattson. 2015. Careful what you share in six seconds: Detecting cyberbullying instances in Vine. In ASONAM. ACM, 617–622.

Python packages version

• seaborn==0.11.1
• torch==1.7.0
• matplotlib==3.3.2
• pandas==1.2.0
• numpy==1.19.2
• ipython==7.20.0
• scikit_learn==0.24.1

Reference

Lu Cheng*, Ahmadreza Mosallanezhad*, Yasin N. Silva, Deborah L. Hall, and Huan Liu. Mitigating Bias in Session-based Cyberbullying Detection: A Non-Compromising Approach. The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021), 2021.