An Extensive Survey on Intrusion Detection- Past, Present, Future

Intrusion Detection is the most eminent fields in the network which can also be called as anomaly detection. Various methods used by early research tells that, the kind of measures used to detect the intrusion is not specified. Research has grown extensively in Anomaly intrusion detection by using different data mining techniques. Most researchers have not briefed on the kinds of distances measures used, the classification and feature selection techniques used in identifying intrusion detection. Intrusion detection is classified with problems as Outlier problems, Sparseness problem and Data Distribution. One of the important observations made is, High Dimensional Data Reduction is not performed, and conventional dataset is not used or maintained by any researchers. A survey is performed to identify the type of distance measures used and the type of datasets used in the early research. In this extended survey, the measures like Distance measure, pattern behaviors are used in identifying the Network Intrusion Detection. In this paper, we present the various methods used by authors to obtain feature selection methods. Also, the discussion is towards, Computation of High Dimensional Data, how to decide the Choice of Learning algorithm, Efficient Distance and similarity measures to identify the intrusion detection from different datasets.