Outlier Detection Based on Temporal Data

Abstract

The objective of research is to provide an efficient method for Network Intrusion newlinedetection. The objective is achieved by a comprehensive process which includes newlinedetail literature survey of outlier detection, methods and applications. Outlier newlinedetection is explored in such a way that provides a sound understanding to achieve newlineefficient network intrusion detection. Recently most of the system uses temporal data newlinein every application. One of the important application areas of temporal outlier newlineanalysis and detection is network anomaly detection. In the last two decades NAD is newlineused by several network administrators to achieve efficient detection to improve the newlinesystem performance. A number of NAD systems are developed with the help several newlinetraditional methods as well as new methods. Every system is used in different newlineenvironment to achieve efficient detection. All of these systems are modeled and newlinetested with some standard network datasets which are available publicly. Due to the newlineincrease in the number of attacks and dependency of every enterprise towards newlinededicated network infrastructure, it is necessary to update NAD system regularly to newlinetackle novel intrusion and sophisticated attacks. A hybrid system with several newlinemachine learning algorithm provide an efficient detection model with better accuracy newlineand performance for network anomaly detection.