An Enhanced Machine Learning Framework for Advanced Analysis and Monitoring of Women’s Safety on Social Media Platforms

The rapid growth of social media has brought about significant negative consequences, such as cyberbullying, cyber abuse, and online trolling, which disproportionately affect women and children, often leading to severe psychological and physical distress. The increasing instances of online harassment, including cyberbullying and abusive messages on platforms like Twitter, Facebook, and Instagram, highlight the urgent need for protective measures, as these can escalate to serious outcomes, including suicidal tendencies. While societal organizations like She Team and Disha Act aim to protect women, there is a lack of robust measures and frameworks on social media platforms to safeguard women from online threats and harassment. Examining the frequency and kind of threats, as well as the severity of online harassment, this research seeks to investigate the prevalence of abusive material directed against women on different social media platforms in urban areas of India. Research on the security of women on different social media platforms makes use of algorithms like support vector machines (SVMs), random forests (RFs), and neurons bayes (NB). Classification methods will be used in order to classify or predict the Type according to attributes in the dataset. We can find out whether social media posts are favorable, bad, or neutral by using classification algorithms. As opposed to support vector machines (SVMs) and Naive Bayes algorithms, the Random Forest model demonstrated superior accuracy, reaching an estimated 89.53%.