Machine Learning Approaches to Cyber Security
Shantanu Prasad Burnwal · Research Archive of Indian Institute of Technology Hyderabad (Indian Institute of Technology Hyderabad) · 2016
Cyber-security is used to identify cyber-attacks while they are acting on a computer or network system to compromise security of the system. We discuss the concept of Hidden Markov Model with the Large Deviation Theory approaches because now a days statistical anomaly detection with Large Deviation theory approach have been used to find attack signatures in network traffic. We present two different approaches to characterize traffic: a model-free approach and a model-based approach. Model free approach is method of types based approach using Sanov’s theorem whereas model based approach is HMM based approach uses Large deviation theory. We discuss how these theories can be applied for anomaly detection from network traffic. We study their effectiveness in anomaly detection. We will discuss how much these statistical methods affective on spatio-temporal traffic data. We also discuss about how length of traffic data affect our Markov model. How our estimated model is related with true but unknown model.