A Survey of Offensive Language Detection for the Arabic Language

Abstract

The use of offensive language in user-generated content is a serious problem that needs to be addressed with the latest technology. The field of Natural Language Processing (NLP) can support the automatic detection of offensive language. In this survey, we review previous NLP studies that cover Arabic offensive language detection. This survey investigates the state-of-the-art in offensive language detection for the Arabic language, providing a structured overview of previous approaches, including core techniques, tools, resources, methods, and main features used. This work also discusses the limitations and gaps of the previous studies. Findings from this survey emphasize the importance of investing further effort in detecting Arabic offensive language, including the development of benchmark resources and the invention of novel preprocessing and feature extraction techniques.

Detecting Arabic textual threats in social media using artificial intelligence: An overview

“…Actually, personal knowledge and culture are both crucial elements in defining what is offensive and what is not. A word used in a written or spoken communication is offensive if it contains conduct meant to cause hurt, pain, or anger [10], [16]. Hate speech and cyberbullying are two significant forms of offensive language, their prevalence on social media has recently risen.…”

“…It is described as "any use of modern digital technology to propagate racial, religious, extremist, or terrorist ideas" [18]. Hate speech can be classified into the following categories: gendered and religious [10], [19].…”

“…Advanced technical approaches that can aid social media platforms and others in implementing these laws [6]. Online offensive language spotting has been used in multiple languages, such as English, German, Turkish, Hindi, Chinese, and Arabic [7]- [10].…”

The Modern Greek Language on the Social Web: A Survey of Data Sets and Mining Applications

“…Binary SVM classifiers have employed textual content as feature vectors by implementing feature selection on each word (uni-grams). More specifically, they applied the method dictated by the bag-of-words language model 32 , and every individual is considered as a single word feature. Each feature cumulatively forms the data set, which represents the corpus consisting of a document per column matrix.…”

“…Therefore, linguistic idiom is a feature of Greek social text that needs to be studied separately. Nevertheless, researchers have studied the reaction patterns of social media users in the language of their interest, such as Czech [31], Arabic [32], as well as in different languages [33] on various applications, including sentiment analysis, regarding either a single language or a multilingual setting [19,20,34].…”