Advancements in Machine Learning Technologies for Detecting and Addressing Harmful Behaviours in Online Platforms

Abstract

Abstract: Social engineering attacks, which take use of psychological manipulation rather than technical flaws, represent a serious threat to people, organizations, and society at large. This research paper explores the field of social engineering detection and provides a thorough analysis of current approaches, their efficacy, and the difficulties encountered in spotting and countering these misleading techniques. This paper presents a comprehensive overview of the use of machine learning and deep learning technologies to detect and address harmful behaviours such as hate speech, cyberbullying, phishing, and spam on social media and communication platforms. It covers a range of studies and methods, including the application of various algorithms such as Enigma Cipher, Rail Fence Cipher, Convolutional Neural Networks, BiLSTM, decision trees, Logistic Regression, SVM, Naive Bayes, and others. The document highlights ongoing endeavours to tackle the issues posed by harmful activities online through innovative technological advancements, showcasing the continuous evolution in this dynamic field.