Intelligent detection of hate speech in Arabic social network: A machine learning approach

Abstract

Nowadays, cyber hate speech is increasingly growing, which forms a serious problem worldwide by threatening the cohesion of civil societies. Hate speech relates to using expressions or phrases that are violent, offensive or insulting for a person or a minority of people. In particular, in the Arab region, the number of Arab social media users is growing rapidly, which is accompanied with high increasing rate of cyber hate speech. This drew our attention to aspire healthy online environments that are free of hatred and discrimination. Therefore, this article aims to detect cyber hate speech based on Arabic context over Twitter platform, by applying Natural Language Processing (NLP) techniques, and machine learning methods. The article considers a set of tweets related to racism, journalism, sports orientation, terrorism and Islam. Several types of features and emotions are extracted and arranged in 15 different combinations of data. The processed dataset is experimented using Support Vector Machine (SVM), Naive Bayes (NB), Decision Tree (DT) and Random Forest (RF), in which RF with the feature set of Term Frequency-Inverse Document Frequency (TF-IDF) and profile-related features achieves the best results. Furthermore, a feature importance analysis is conducted based on RF classifier in order to quantify the predictive ability of features in regard to the hate class.

Intelligent Detection of False Information in Arabic Tweets Utilizing Hybrid Harris Hawks Based Feature Selection and Machine Learning Models

“…Text vectorization, or more simply, feature extraction, is the process of converting the unstructured text into structured representation (numerical features) such that ML algorithms can be applied for mining and knowledge extraction purposes [2]. In simple terms, the numerical features are extracted utilizing words-based statistical measurements.…”

“…Therefore, the text is viewed as a multivariate sample (vector) of such measures. Bag of Words (BoW) is a common text representative model that is widely utilized for text classification applications [2,12,31]. It is a flexible and straightforward textual representation scheme that describes the occurrences or count of terms within a piece of text without considering the exact ordering and semantic structure.…”

“…Various evaluation measures can be derived from the outcomes of the confusion matrix. In this study, the used measures are accuracy, recall, precision, and F1_measure which are widely used in the text classification domain [2,12,31]. Details and mathematical expression of these evaluation measures are listed as follows:…”

“…In the social media mining field, data mining and text mining techniques are exploited in conjunction with each of social network analysis and information retrieval for gaining useful hidden data that uncover the purpose of spreading news, sentiments, or opinions over social media platforms [1,2]. Understanding or monitoring the sentiment tendency of people is essential for making the proper decision in organizations.…”

Automatic Classification of Sexism in Social Networks: An Empirical Study on Twitter Data

“…In this task, deep learning approaches were proposed to detect misogyny and racism in texts in Spanish [34]. Similarly, [37] aims to detect cyber hate speech in Arabic tweets employing a wide range of traditional machine learning techniques.…”