Digital Media Criminal Justice: The study Innovative Neural Network-Based Hateful Communication Recognition System Associated with Online Harassment Without Unpredictability

Computer science

Cyber bullying emerged as a critical issue on social media platforms, posing a global threat not only to victims but also to the broader society. Automatically detecting cyber bullying remains a formidable challenge due to the complex language, subtlety, and informal nature of communication on these platforms. The brevity and ambiguity of social media content often obscure the intended meaning, making it difficult to interpret and classify accurately. This complexity is further intensified when content lacks clear context or exhibits vague or uncertain expressions. Although several existing methods aim to detect cyber bullying, they often struggle to differentiate between various forms of abusive language, especially due to the nuanced and overlapping nature of hate speech. In response to these challenges, This proposed method is designed to enhance the identification of different cyberbullying categories by addressing ambiguity and the fuzzy boundaries between distinct classes. Neutrosophic Logic is employed to manage uncertainty, vagueness, and indeterminacy within the classification process, enabling a more adaptable and the robust framework for handling complex textual scenarios. The model utilizes a one-against-one classification strategy within the MLP to effectively capture and differentiate the intricate relationships among various cyber bullying types, especially in cases where categories overlap or are not clearly defined.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.