Cyberbullying on social media, especially in the context of political issues, has become a significant concern in Bangladesh. Researchers have recently developed a machine learning model to detect cyberbullying in Bangla text on social media, achieving an accuracy rate of 91.08%.
The study, published in the "Journal of Intelligent Learning Systems and Applications," focused on the Bangla language, which is widely used for political discourse on social media in Bangladesh.
To develop their model, the researchers collected a dataset of 11,000 Bangla texts from the comments section of political Facebook posts and manually labeled them as either bullying or not. This dataset was used to train various machine learning classifiers.
The results showed that the Random Forest classifier outperformed other algorithms, achieving an accuracy rate of 91.08%. This classifier also demonstrated high precision (0.78), recall (0.82), and F1-score (0.80). The area under the Receiver Operating Characteristic (ROC) curve was also larger for the Random Forest classifier, indicating its effectiveness in identifying instances of political bullying in Bangla text.
The research addresses the lack of annotated corpora and morphological analyzers for Bangla text and provides a promising solution for monitoring and detecting cyberbullying on social media, especially in the context of political issues.
This work not only contributes to the understanding of cyberbullying in Bangla text but also offers insights into machine-learning approaches for detecting and preventing online harassment in different languages and regions.
Share This Post On
Leave a comment
You need to login to leave a comment. Log-in