A Unified System for Aggression Identification in English Code-Mixed and Uni-Lingual Texts

Abstract

Wide usage of social media platforms has increased the risk of aggression, which results in mental stress and affects the lives of people negatively like psychological agony, fighting behavior, and disrespect to others. Majority of such conversations contains codemixed languages [28]. Additionally, the way used to express thought or communication style also changes from one social media platform to another platform (e.g., communication styles are different in twitter and Facebook). These all have increased the complexity of the problem. To solve these problems, we have introduced a unified and robust multi-modal deep learning architecture which works for English code-mixed dataset and uni-lingual English dataset both. The devised system, uses psycho-linguistic features and very basic linguistic features. Our multi-modal deep learning architecture contains, Deep Pyramid CNN, Pooled BiLSTM, and Disconnected RNN(with Glove and FastText embedding, both). Finally, the system takes the decision based on model averaging. We evaluated our system on English Code-Mixed TRAC 1 2018 dataset and uni-lingual English dataset obtained from Kaggle 2 . Experimental results show that our proposed system outperforms all the previous approaches on English code-mixed dataset and uni-lingual English dataset.

You are what your feeds makes you: A study of user aggressive behaviour on Twitter

“…These features are optimized using binary particle swarm optimization (BPSO) and binary firefly optimization (BFFO) algorithms. Khandelwal and Kumar (2020) proposed a multimodal deep learning architecture with linguistic and psychological linguistic features for aggression detection in code-mixed conversations.…”

“…Despite an upsurge in research on aggressive behavior on Twitter, the relationship between consuming aggressive content and user behavior remains unexplored. Previous research has primarily focused on aggression detection on social media platforms using various machine learning (Datta et al, 2020;Arroyo-Fernández et al, 2018;Gutiérrez-Esparza et al, 2019), deep learning models (Aroyehun and Gelbukh, 2018;Srivastava and Khurana, 2019;Kumari et al, 2021), as well as identifying aggression in multilingual (Sharif and Hoque, 2022;Kumari et al, 2021;Torregrosa et al, 2022) and multi-modal posts (Khandelwal and Kumar, 2020;Kumari and Singh, 2022) on social media. Apart from aggression detection, some sociological studies of aggressive behavior on social media exist (Vladimirou et al, 2021;Pascual-Ferrá et al, 2021).…”