Hate Speech in the Arab Electronic Press and Social Networks

Abstract

Nowadays we are witnessing an open world, characterized by globalization which is accompanied by a technology through which information circulates without borders, especially with the widespread use of social networking sites being the most common communication tool, that gives access through various applications to a large space for the presentation of multiple ideas, including extremist ideas, and the spread of hate speech. This paper introduces a system of detection of hate speech in the texts of Arabic read media and social media, which is based on a combined use of NLP, and machine learning methods. The training of the detection model is done on a large Dataset of articles, tweets and comments, collected, balanced and tokenized afterwards using BERT in Arabic. The trained model detects hate speech in Arabic and various Arabic based dialects, by classifying the texts into two classes: Neutral and Abusive. The above-mentioned model is evaluated using precision metrics, recall and f1 score, it has reached an accuracy of 83%.

Toxic language detection: A systematic review of Arabic datasets

“…Besides, three datasets are extracted from the web, which are the large‐scale corpus OSCAR, where adult content is labelled (Abadji et al, 2022; Jansen et al, 2022), the dataset of the deleted comments from the website of the news channel Aljazeera (Mubarak et al, 2017) and Riabi et al (2023) dataset, which includes user comments from a news website in addition to texts from a corpus of song lyrics. Several datasets are multi‐sources (Alduailaj & Belghith, 2023; Awane et al, 2021; Badri et al, 2022; Bensalem et al, 2023; Chowdhury et al, 2020; M. Habash & Daqour, 2022; Haddad et al, 2019; Khairy et al, 2022; Mazari & Kheddar, 2023; Riabi et al, 2023), which explains their presence with more than one source in Table 4.…”

“…In (Fraiwan, 2022), for example, the published dataset contains more than 200 additional examples compared to the dataset used in experiments as described in the paper. An extreme case is the work of (Awane et al, 2021), where there is no correspondence between the dataset and the paper, not only in the number of examples but also in the annotation labels and the ratio of offensive examples.…”

“…Our examination revealed certain cases where the data statistics described in the research paper do not perfectly match the dataset available online, as observed in (Alduailaj & Belghith, 2023;Alshalan & Al-Khalifa, 2020;Awane et al, 2021;Fraiwan, 2022). In (Fraiwan, 2022), for example, the published dataset contains more than 200 additional examples compared to the dataset used in experiments as described in the paper.…”

“…Alakrot et al, 2018) rbz(Albadi et al, 2018) (De Smedt et al, 2018) (Albadi et al, 2019) (Ousidhoum et al, 2019) (Alshehri et al, 2020) (Chowdhury et al, 2020) (Zampieri et al, 2020) ) (Aldera et al, 2021) (Awane et al, 2021) (Hadj Ameur & Aliane, 2021)(Mubarak et al, 2021) (Mubarak, Rashed, et al, 2020) (Mulki & Ghanem, 2021a)(Nakov et al, 2021) (Shaar et al, 2021) (Abadji et al, 2022) (Jansen et al, 2022) (Albadi et al, 2022) (Al Bayari & Abdallah, 2022)(Almanea & Poesio, 2022) (Badri et al, 2022) (Fraiwan, 2022) (Habash & Daqour, 2022) (Khairy et al, 2022)(Mubarak, Hassan, & Chowdhury, 2022) (Nakov et al, 2022) (Obeidat et al, 2022) (Shannag et al, 2022)(Alduailaj & Belghith, 2023) (Bensalem et al, 2023) rbz …”

Hate Speech in the Comments’ Column Instagram: A Discourse Analysis

“…The newly created datasets on offensive content, hate speech, and cyberbullying cover a wide range of non-English languages such as Arabic (Aldjanabi et al, 2021;Aljarah et al, 2021;Alshalan et al, 2020;Awane et al, 2021), Polish (Adamczak-Krysztofowicz & Szczepaniak-Kozak, 2017Habrat, 2021), Spanish (Arango et al, 2020;García-Díaz et al, 2022;Uzan & HaCohen-Kerner, 2021), Italian (Caiani et al, 2021;Celli et al, 2021;Florio et al, 2021), and Chinese (Chen, 2022;Uyheng et al, 2022;Wang et al, 2022). Therefore, it would be fruitful to examine social media as a forum for the propagation of hate speech.…”