Harnessing Machine Learning and Deep Learning for Crime Detection in Social Media through Advanced Data Mining Techniques

Abstract

An enormous quantity of user-generated data is created daily as a result of the exponential development in social media usage. This data includes patterns of illegal activity in addition to being a useful source of public opinion and societal trends. The goal of this work is to create a comprehensive framework that uses data mining, deep learning, and machine learning (ML) to identify illegal activity using social media analytics. A model that recognizes suspicious conduct, hate speech, terrorist communication, and cyberbullying is proposed and validated in the study. Using both supervised and unsupervised learning techniques, experiments were performed on datasets taken from Reddit and Twitter. When compared to current approaches, the suggested strategy showed encouraging outcomes in terms of accuracy, recall, and F1-score.