A network intrusion prediction model using Bayesian network

Intrusion detection systems are increasingly becoming more and more useful in the fight against cyber threats. As businesses continue to transition into adopting information processing systems including cloud computing, there is a greater need for the development of safety measures to ensure the preservation of information against theft, unauthorized access and other cyber threats. Several articles in literature have reported the development of intrusion detection and prediction system using varying techniques. In this research work, a simple Bayesian model is presented. To build this model, the widely used dataset of NSL-KDD was used. Existing literature was relied on for the selection of the best features while the Max-Min Hill Climbing (MMHC) algorithm was used to define the Bayesian network structure. The model was then implemented using the Genie Bayesian Modelling software. Results of simulation, influence and scenario analysis established the efficacy of the model in predicting the likelihood of an intrusion in a network environment.