Nlp Driven Cyberbullying Detection System for Social Media Using Transformed Based Sentimental Analysis
Year: 2026
Type: article
Abstract: The rapid growth of social media platforms hassignificantly transformed digital communication byenabling users to share information, opinions, andmultimedia content globally. However, thiswidespread connectivity has also increased theoccurrence of cyberbullying, online harassment,abusive comments, and toxic interactions thatnegatively affect individuals, particularly teenagersand young users. Cyberbullying can lead to severepsychological, emotional, and social consequencesincluding anxiety, depression, stress, and suicidaltendencies. Manual monitoring of harmful contentacross millions of social media posts is highlyinefficient and impractical due to the enormousvolume of user-generated data. Therefore, anautomated and intelligent cyberbullying detectionsystem is essential for ensuring safer onlinecommunication. This project presents an NLPDriven Cyberbullying Detection System for SocialMedia Using Transformer Based SentimentalAnalysis that combines Natural LanguageProcessing techniques with machine learningalgorithms to identify abusive and bullying contenteffectively. The system performs text preprocessingoperations such as tokenization, stop-wordremoval, normalization, and stemming to cleannoisy social media data. Feature extractiontechniques including Bag-of-Words and TF-IDF areapplied to convert textual information intonumerical vectors suitable for classification.Transformer-based sentiment analysis isincorporated to capture contextual meaning andsemantic relationships within messages, improvingthe detection of implicit and sarcastic bullyingcontent. Multiple machine learning classifiers suchas Support Vector Machine, Logistic Regression,and Naive Bayes are evaluated to achieve highclassification accuracy. The proposed systemprovides automated monitoring, scalabledeployment, improved text representation, andenhanced detection performance for real-timecyberbullying identification. The system ultimatelycontributes to creating a safer digital environmentand supports social media moderation throughintelligent harmful content filtering.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7216775264