Enhancing the Efficiency of Attack Detection System Using Feature selection and Feature Discretization Methods

Intrusion detection technologies have grown in popularity recent years using machine learning. The variety of new security attacks are increasing, necessitating the development effective and intelligent countermeasures. existing intrusion system (IDS) uses Signature or Anomaly based systems with learning algorithms to detect malicious activities. Signature-based rely only on signatures that been pre-programmed into systems, known cannot any unusual activity. supervised algorithm detects threats. To address this issue, proposed model employs an unsupervised approach for detecting attacks. This combines Sub Space Clustering One Class Support Vector Machine utilizes feature selection methods such as Chi-square, well Feature Discretization Methods like Equal Width identify both undiscovered assaults. results experiments outperforms several terms rate accuracy decrease computational time.