A network attack discovery algorithm based on unbalanced sampling vehicle evolution strategy for intrusion detection

ABSTRACT

Aiming at the problem of mass data processing in intrusion detection model, a kind of intrusion detection model algorithm based on minimum rule self-organizing map (SOM) elliptical fuzzy patch projection membership function (NF) is proposed in order to better reduce the computation complexity of algorithm and improve the detection accuracy. For the condition that a large number of SOM nodes will lead to complex model and over-fitting model, the problem identification of optimal SOM grid size shall be handled by Person Correlation Coefficient to give the best determination suggestion for SOM grid size. And then the construction method of elliptical fuzzy patch based on the improved Gauss membership function estimation and the calculation method of improved Gauss membership function of relevant data are provided. Finally, the results of simulation comparison for true intrusion detection data-set show that, the proposed method is superior to selected comparison algorithm in detecting accuracy and calculating time, verified the effectiveness of proposed method.

KEYWORDS:

• Self-organizing map
• differential evolution
• fuzzy patch
• intrusion detection
• network attack
• classification feature

Additional information

Funding

In 2013, The Ministry of Education of Guangdong Province made a major achievements transformation project [item number: 2013B090200055].