This paper represents a converged exposition of machine learning approach on various network security applications specifying the possible future scope and detailed outcome of the researches on the basis of intrusion detection, analysis of malware behavior, internet traffic classification, and other security aspects. The complexity of the respective algorithms and the challenges of using them for network security applications are also discussed which would help in understanding the current scenario and the future prospects.