Bayesian theory and association rule mining methods are artificial intelligence techniques that have been used in various computing fields, especially in machine learning. Internet has been considered as an easy ground for vices like radicalization because of its diverse nature and ease of information access. These vices could be managed using recommender systems methods which are used to deliver users’ preference data based on their previous interests and in relation with the community around the user. The recommender systems are divided into two broad categories, i.e. collaborative systems which considers users which share the same preferences as the user in question and content-based recommender systems tends to recommend websites similar to those already liked by the user. Recent research and information from security organs indicate that, online radicalization has been growing at an alarming rate. The paper reviews in depth what has been carried out in recommender systems and looks at how these methods could be combined to from a strong system to monitor and manage online menace as a result of radicalization. The relationship between different websites and the trend from continuous access of these websites forms the basis for probabilistic reasoning in understanding the users’ behavior. Association rule mining method has been widely used in recommender systems in profiling and generating users’ preferences. To add probabilistic reasoning considering internet magnitude and more so in social media, Bayesian theory is incorporated. Combination of this two techniques provides better analysis of the results thereby adding reliability and knowledge to the results.