Hate and offensive speech detection on Arabic social media

Alsafari, Safa; Sadaoui, Samira; Mouhoub, Malek

doi:10.1016/j.osnem.2020.100096

Cited by 67 publications

(54 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Both methods are used to learn the usage context of a word. Global vector space [20], [25], [26], [27], [28], [29], [30] Distance Measures Edit-Distance [20] Word Embedding Word2Vec [31], [32], [28], [33], [34], [35], [30], [36], [37], [38], [39] Skip-gram [25] CBoW [25], [32] BoW [40], [31], [33], [34] TF-IDF [26], [27] FastText [25], [36] GLoVe [41], [42], [37] LSHWE [37] Vulgarity/Hate Features [43], [25], [32], [33], [44] Sentiment Sentiment Analysis [27], [45], [32], [33], [41], [46], [30], [39] User Profile [27],…”

Section: Word Embedding Techniquesmentioning

confidence: 99%

A Systematic Review of Machine Learning Algorithms in Cyberbullying Detection: Future Directions and Challenges

Arif

2021

JISCR

View full text Add to dashboard Cite

Social media networks are becoming an essential part of life for most of the world’s population. Detecting cyberbullying using machine learning and natural language processing algorithms is getting the attention of researchers. There is a growing need for automatic detection and mitigation of cyberbullying events on social media. In this study, research directions and the theoretical foundation in this area are investigated. A systematic review of the current state-of-the-art research in this area is conducted. A framework considering all possible actors in the cyberbullying event must be designed, including various aspects of cyberbullying and its effect on the participating actors. Furthermore, future directions and challenges are also discussed.

show abstract

Section: Word Embedding Techniquesmentioning

confidence: 99%

A Systematic Review of Machine Learning Algorithms in Cyberbullying Detection: Future Directions and Challenges

Arif

2021

JISCR

View full text Add to dashboard Cite

show abstract

“…In a recent study, (Alsafari, Sadaoui, and Mouhoub 2020c) built a robust hate speech corpus written in the two common Arabic languages: Modern Standard Arabic, which is understandable by all Arabic speakers, and the Gulf Arabic dialect, which is spoken in the countries of the Arabian Peninsula. The authors first queried the Twitter platform using four different searching strategies: keyword, hashtag, profile, and defensive methods.…”

Section: Labeled and Unlabeled Corporamentioning

confidence: 99%

“…This Arabic corpus for hate speech classification was evaluated extensively using supervised deep learning algorithms combined with various text vectorization methods (Alsafari, Sadaoui, and Mouhoub 2020c;2020a;2020b). Therefore, we consider this corpus reliable and use it for the initial ESSL phase.…”

Section: Labeled and Unlabeled Corporamentioning

confidence: 99%

“…Past studies on hate speech detection have adopted supervised learning across all the examined languages, including English (Zampieri 2020;Zampieri et al 2019) and Arabic (Alsafari, Sadaoui, and Mouhoub 2020c;Mubarak et al 2020). These studies were conducted with small corpora because obtaining labeled data is costly and time-consuming.…”

Section: Introductionmentioning

confidence: 99%

“…for hate speech detection; this study was conducted for the English language and was based on co-training two classifiers (Rosenthal et al 2020). We evaluate our approach's performance using a robust labeled Arabic dataset (publicly available) developed and tested by (Alsafari, Sadaoui, and Mouhoub 2020c) and a massive unlabeled dataset of one million tweets that we scraped from Twitter, cleaned, and normalized. As baseline models for the ESSL process, we develop three diverse hate speech classifiers: SVM with Word and Char Ngram, CNN, and BiLSTM; the latter two were both merged with W2V Skipgram Word embeddings.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Ensemble-based Semi-Supervised Learning for Hate Speech Detection

Alsafari

Sadaoui

2021

FLAIRS

Self Cite

View full text Add to dashboard Cite

Large and accurately labeled textual corpora are vital to developing efficient hate speech classifiers. This paper introduces an ensemble-based semi-supervised learning approach to leverage the availability of abundant social media content. Starting with a reliable hate speech dataset, we train and test diverse classifiers that are then used to label a corpus of one million tweets. Next, we investigate several strategies to select the most confident labels from the obtained pseudo labels. We assess these strategies by re-training all the classifiers with the seed dataset augmented with the trusted pseudo-labeled data. Finally, we demonstrate that our approach improves classification performance over supervised hate speech classification methods.

show abstract