Eliminate False Alerts in Intrusion Detection using Deep Learning

Abstract

With Advance in technology and Internet, security of Personal information and Sys- tem (computer) is becoming a major problem.As time is going, numbers of attacks on systems are increasing. Intrusion detection plays major role in identifying security is- sues.However, there are certain Limitations of Intrusion Detection System.One of them is False alarm. Meaning of false alarm is, it ags normal behaviour as Intrusion. Intrusion detection system generates large amount of false alarm. To overcome limitations ,previ- ous researcher have used machine learning algorithms like Support vector machine and K-nearest neighbours.In this paper, I am using Deep belief network and self organizing map to eliminate false alarm. At last, this paper represent performance of deep learning approach with previous work. Comparison of diffierent approaches are based on accuracy, f-score , precision and recall.