![](bg8.png)
8 CONCLUSION
In this study, we explore a strategy to detect hate-
speech. We based our approach on considering mes-
sages from several online social-media platforms at
once, betting that their different internal moderation
policies would provide a larger set of haters’ meth-
ods. In additions to sharing our annotated set with
the community, we also develop an application build-
ing up our strategy of combining/comparing multiple
pairs of word embeddings and classifiers. Overall,
we successfully build a hate speech detection model,
pairing USE and SVC, to obtain an average accu-
racy of 95.65% and achieved a maximum accuracy of
96.89%. Moreover, our application allows to define
an aggregating strategy by e.g. choosing which pairs
should be taken more into account. Therefore, we
hope that this two-side strategy of involving several
platforms and combining multiple pairs of embed-
dings and classifiers, will inspire the community to
improve our results and refine our performance score.
Sensitive Content Warning. Due to the nature of
this study, there are places in this article where hate-
ful language and terms are used. While we did try and
keep the use of these terms and phrases to a minimum,
and while we obviously do not approve these mes-
sage, it was vital to provide the reader with a proper
understanding of the context and methodologies used
in the process of completing this project.
REFERENCES
Alshalan, R. and Al-Khalifa, H. (2020). A deep learning
approach for automatic hate speech detection in the
saudi twittersphere. MDPI.
Bhattacharya, D. and Weber, I. (2019). Racial bias in hate
speech and abusive language detection datasets. ACL
Anthology.
Cer, D., Yang, Y., Kong, S.-y., Hua, N., Limtiaco, N., John,
R. S., Constant, N., Guajardo-Cespedes, M., Yuan,
S., Tar, C., et al. (2018). Universal sentence encoder.
arXiv preprint arXiv:1803.11175.
Chen, T. and Guestrin, C. (2016). Xgboost: A scalable
tree boosting system. In Proceedings of the 22nd acm
sigkdd international conference on knowledge discov-
ery and data mining, pages 785–794.
Davidson, T., Warmsley, D., Macy, M., and Weber, I.
(2017). Automated hate speech detection and the
problem of offensive language.
Devlin, J., Chang, M.-W., Lee, K., and Toutanova,
K. (2018). Bert: Pre-training of deep bidi-
rectional transformers for language understanding.
arXiv:1810.04805.
Freund, Y. and Schapire, R. E. (1997). A decision-theoretic
generalization of on-line learning and an application
to boosting. Journal of computer and system sciences,
55(1):119–139.
Gibbs, M. N. (1998). Bayesian Gaussian processes for re-
gression and classification. PhD thesis, Citeseer.
Guo, G., Wang, H., Bell, D., Bi, Y., and Greer, K. (2003).
Knn model-based approach in classification. In OTM
Confederated Int. Conf., pages 986–996. Springer.
Han, E. (2020). Countering hate on tiktok.
Hawdon, J., Oksanen, A., and R
¨
as
¨
anen, P. (2015). Online
extremism and online hate exposure among adoles-
cents and young adults in four nations.
Hornik, K., Stinchcombe, M., and White, H. (1989). Multi-
layer feedforward networks are universal approxima-
tors. Neural networks, 2(5):359–366.
Izenman, A. J. (2013). Linear discriminant analysis.
In Modern multivariate statistical techniques, pages
237–280. Springer.
Jourov
´
a, V. (2016). Code of conduct - illegal online hate
speech questions and answers.
Kansara, K. and Shekokar, N. (2015). A framework for
cyberbullying detection in social network. Semantic
Scholar.
Kov
´
acs, G., Alonso, P., and Saini, R. (2020). Challenges of
hate speech detection in social media. Springer Na-
ture.
L1GHT (2020). Rising levels of hate speech & online toxi-
city during this time of crisis.
Le, Q. and Mikolov, T. (2014). Distributed representations
of sentences and documents. In International confer-
ence on machine learning, pages 1188–1196. PMLR.
Lee, H. S. and Lee, H. R. (2018). An abusive text detection
system based on enhanced abusive and non-abusive
word lists. Yonsei University.
Mozafari, M. (2020). Hate speech detection and racial bias
mitigation in social media based on bert model. PLOS
ONE.
Pal, M. (2005). Random forest classifier for remote sensing
classification. International journal of remote sensing,
26(1):217–222.
Ramampiaro, H. (2018). Detecting offensive language in
tweets using deep learning. Cornell University.
Rish, I. et al. (2001). An empirical study of the naive bayes
classifier. In IJCAI 2001 workshop on empirical meth-
ods in artificial intelligence.
Rudnicki, K. and Steiger, S. (2020). Online hate speech.
R
¨
ottger, P., Vidgen, B., Nguyen, D., Waseem, Z., Margetts,
H., and Pierrehumbert, J. (2021). Hatecheck: Func-
tional tests for hate speech detection models.
Safavian, S. R. and Landgrebe, D. (1991). A survey of de-
cision tree classifier methodology. IEEE transactions
on systems, man, and cybernetics, 21(3):660–674.
Salminen, J., Hopf, M., Chowdhury, S., gyo Jung, S.,
Almerekhi, H., and Jansen, B. (2020). Developing an
online hate classifier for multiple social media plat-
forms. Human-centric Computing and Info. Sciences.
ICAART 2023 - 15th International Conference on Agents and Artificial Intelligence
548