CoChat
Sign Up
Cited by 396Hate Speech and Cyberbullying DetectionSentiment Analysis and Opinion MiningMisinformation and Its Impacts

A BERT-Based Transfer Learning Approach for Hate Speech Detection in Online Social Media

Open access
Mozafari, Marzieh, Farahbakhsh, Reza, Crespi, Noël·2020·Studies in Computational Intelligence
DOI 10.1007/978-3-030-36687-2_77
Open this paper in CoChat — free accountOpen PDF

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1007/978-3-030-36687-2_77

Cites

37 of 37 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • BERT has produced outstanding results on natural language understanding tasks, and in numerous studies it outperformed conventional machine learning algorithms and demonstrated promising results in cyberbullying detection.

    “On a variety of natural language understanding tasks, Bidirectional Encoder Representation from Transformers (BERT) has produced outstanding results. In numerous studies, BERT outperformed conventional machine learning algorithms, achieved cutting-edge performance, and demonstrated promising results in the detection of cyberbullying.”

    From A Review on Deep-Learning-Based Cyberbullying Detection (2023) · cites this paper as [60]

    Previous Works (p. 3) reports that BERT outperformed earlier language models on several NLP tasks. The Introduction (p. 2) and Conclusion (p. 10) report that this paper’s BERT-based model outperformed previous approaches to hate-speech detection. The paper does not establish that numerous BERT studies outperformed conventional machine-learning algorithms, and its mention of cyberbullying in Previous Works (p. 3) concerns a different model, not a BERT result. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Mozafari et al. found encouraging experimental results for BERT in detecting cyberbullying in two social media datasets in terms of precision, recall, and F1 scores compared to prior research.

    “To assess the effectiveness of BERT in detecting cyberbullying in a social media context, Mozafari et al. [60] used BERT to categorize cyberbullying in two social media datasets. In terms of precision, recall, and F1 scores, their experimental findings are encouraging when compared to the prior research in this area.”

    From A Review on Deep-Learning-Based Cyberbullying Detection (2023) · cites this paper as [60]

    The abstract (p. 1) says the study used two Twitter datasets annotated for racism, sexism, hate, or offensive content—not cyberbullying. The Introduction (p. 2) and Results (p. 8) report BERT performance against prior work using precision, recall, and F1-score, supporting the comparison but not the claim about what was detected. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Studies show BERT achieved high accuracy and F1 scores to classify cyberbullying in various types of online content, such as tweets and comments.

    “Studies [22,58-60] show that BERT achieved high accuracy and F1 scores to classify cyberbullying in various types of online content, such as tweets and comments.”

    From A Review on Deep-Learning-Based Cyberbullying Detection (2023) · cites this paper as [60]

    The Results Analysis (p. 8) reports strong F1 scores for a BERT-based classifier on two Twitter hate-speech datasets. It reports precision, recall, and F1—not accuracy—and the study does not evaluate cyberbullying classification across content types such as comments. (LLM, conf=0.89)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Mozafari, M., Farahbakhsh, R., & Crespi, N.
Affiliations
Centre National de la Recherche Scientifique (FR)
Corresponding author
Mozafari, Marzieh
ORCID
Venue
Studies in Computational Intelligence
Publisher
Springer International Publishing
Published
2019-11-26
Volume
Studies in Computational Intelligence book series (SCI, volume 881)
Pages
928-940
ISSN
  • 1860-949X (ISSN-L)
  • 1860-9503
DOI
10.1007/978-3-030-36687-2_77
OpenAlex
W2981478815
Cited by
396
References
31
Citations by year
2026: 25 · 2025: 44 · 2024: 67 · 2023: 81 · 2022: 73 · 2021: 75 · 2020: 31
FWCI
17.12
Type
conference paper
Language
en
Open access
Licence
www.springer.com/tdm (tdm)
Locations
  • Studies in computational intelligence (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • HAL (submittedVersion, other-oa)
  • arXiv (Cornell University)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
  • Physical Sciences › Computer Science › Artificial Intelligence › Sentiment Analysis and Opinion Mining
  • Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
Keywords
hate speech detection, BERT, transfer learning, fine-tuning, social media, offensive content, racism, sexism, Twitter

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review