Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{178362,
        author = {G. Kokila and K. Abivarsha and P. Muthuselvam and A. Rahul sarath and M. R. Sneha},
        title = {DETECTION OF CYBER BULLYING ON SOCIAL MEDIA USING ML},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {4280-4286},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=178362},
        abstract = {On social media platforms, cyberbullying has become a widespread and alarming problem that affects people's mental health and general wellbeing all around the world. This paper suggests a cyberbullying detection system that makes use of the Support Vector Machine (SVM) algorithm in order to tackle this issue. The technology seeks to automatically detect and flag instances of cyberbullying in real-time social media content by utilizing machine learning. The first step in creating the detection system is gathering and classifying a large dataset of posts and comments that include instances of both cyberbullying and non-cyberbullying. The bag-of-words or TF-IDF approaches are used to extract important features from the text data after it has been pre-processed by eliminating unnecessary information, tokenizing it, and changing the text to lowercase. The SVM classifier, which looks for the best hyper plane to efficiently separate cyberbullying from non-cyberbullying content, is trained using these converted feature vectors as inputs. Metrics like accuracy, precision, recall, F1-score, and ROC-AUC are analysed to see how well the SVM model performs in detecting instances of cyberbullying using a different testing dataset. To improve the system's performance, model fine-tuning is done by experimenting with different SVM hyper parameters and cross-validation strategies.},
        keywords = {cyberbullying detection, distil Bert, machine learning, pre trained language models},
        month = {May},
        }