A mathematical Model for Selecting Features of Flooding Attacks Detection Methods based on Stability Criterion

In this era, the Internet represents the main tool of data searching and exchanging. By using Internet many apps in different locations, can be connected to each other remotely via web server connectivity. This connectivity is exposed to the internet vulnerabilities such as flooding attacks. Flooding attack is an example of an attack that causes harmful action against the web server resources or network bandwidth. Multivariate Correlation Analysis based Detection approach (MADM) is one of the statistical based detection (NIDS) approaches used to detect such attacks. However, MADM use predefined features that contain some intercorrelated features that have direct impact on classification performance. In this paper, a new extension on MADM architecture is proposed by adding SVM based feature selection method. So the mathematical formulation is explained. The proposed features selection method will be contributing in minimizing the false positive alarm rate and classification error rate.