Over focus on filter methods for IDS. Motivation:
the past decades, Internet and computer systems have raised numerous security
issues due to the explosive use of networks. Any malicious intrusion or attack
on the network may give rise to serious disasters. Intrusion is a malicious,
harmful entity which is responsible for network attack. They violate integrity,
confidentiality and availability of a system resource. In this case, system is failed
to respond for data stolen or being lost. So, Intrusion Detection Systems
(IDSs) are must to decrease the serious influence of these attacks. Intrusion
Detection System is defined as the system or software tool to detect
unauthorized access to a network or computer system.
is capable of detecting all types attack like malicious, harmful attack,
vulnerability, data driven attacks, host based attacks for example privilege
violation, sensitive file access, unauthorized logins and malwares. Then need
IDS once have firewall because the networks having firewall were not designed
to detect attack at network layer and application layer such as worms, viruses,
Denial of services (DoS), distributed denial of services (DDoS) and Trojans.
The work of firewall is to stop external traffic from entering in the internal
intrusions are like viruses, worms, Trojans, or network attacks like
unauthorized login, access of sensitive files, or data driven attacks on
application. The intrusion violates the integrity, confidentiality and
availability. Because of this system is unable to respond or access is denied.
Thus intrusion detection means detection of unauthorized use of system or an
attack on a system or network. The Intrusion detection system (IDS) is a
hardware or software tool to detect these activities. Feature selection is a
technique for eliminating irrelevant and redundant features and selecting the
most optimal subset of features that produce a better characterization of
patterns belonging to different classes. Methods for feature selection are
generally classified into filter and wrapper methods. This paper focus on
filter methods for IDS.
Detection accuracy of anomaly system should maximize.
false positive rate should minimum.
detector generation time should less.
The goal of proposed LSSVM classification algorithm is to
maximize performance in terms of classification accuracy, detection rate, false
positive rate and F-measure.
The objectives of the proposed
application are as follows:
study existing Network Intrusion Detection Systems (NIDSs) and types of NIDSs.
study various machine learning algorithms like Bayesian networks, Neural
networks, fuzzy logic, outlier detection and genetic algorithm.
study current filter-based feature selection approach for detection of
intrusion attacks using Flexible mutual information based feature selection
(FMIFS) and Flexible Linear Correlation Coefficient based Feature Selection
analyze the experimental results of proposed LSSVM+FMIFS and LSSVM+FLCFS
algorithms for intrusion detection system.
propose a new integrated approach for network intrusion detection using JRipper
classifier algorithm, to reduce detector generation time and also to increase
its adaptability and flexibility the studied parameter value selected
automatically according to the used training dataset.
compare the experimental results of LSSVM and Ripper classifier for network
intrusion detection algorithm which gives results as accuracy of classification