Computational CBGSA – SVM Model for Network Based Intrusion Detection System
The accelerated growth and use of internet-based applications drive the research community towards the development of an appropriate IDS that solely focuses on safeguarding the computer networks from attacks and intrusions by hybridizing various machine learning and statistical intelligent techniques. On considering the open research challenge of designing a robust IDS, this paper proposes an improved Support Vector Machine (SVM) hybridized with Crossover based Binary Gravitational Search Algorithm (CBGSA) for parameter optimization and feature selection. Moreover, the hindrance of local maxima is neglected by introducing the crossover operator during the computation. To illustrate the efficiency of the solution proposed in this paper, CBGSA-SVM has been validated using NSL KDD dataset with the scenarios as follows (i) Training SVM with all the features, and (ii) Training SVM with the optimal feature subset and parameters derived from the CBGSA in terms of detection rate and false alarm rate.
KeywordsSupport Vector Machine Crossover based Binary Gravitational Search Algorithm Intrusion Detection System
- 1.80 Eye-Opening Cyber Security Statistics for 2019. https://www.thesslstore.com/blog/80-eye-opening-cyber-security-statisticsfor-2019/
- 2.Yahoo attack exposes web weakness. http://news.bbc.co.uk/2/hi/science/nature/635444.stm