Improving the accuracy of the intrusion detection system in the detection of DoS attacks using Naive Bayes and comparison with SVM

• Published in: AIP conference proceedings Vol. 2871; no. 1
• Main Authors: Rakesh, T., Nagalakshmi, T. J.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 13.09.2024
• Subjects: Intrusion detection systems; Samples; Statistical analysis; Support vector machines
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0228012

In this study, design of an Innovative Intrusion Detection System using Naive Bayes (Group 2) as compared to the efficiency of SVM, (Group 1). Naive Bayes and Support Vector Machines were the two intrusion detection systems (IDS) used by the researchers. The research uses a total of 38 samples for evaluation, including 19 samples in each group. We used the SPSS program for statistical analysis. A sample size of 80% was determined by combining G power with the pretest power. Using SVM to improve IDS accuracy by a certain percentage is a statistically significant improvement (p = 0.007). The typical level of precision of an Innovative Intrusion Detection System using Naive Bayes (Group 2) is 62.26% and with SVM (Group 1) is 98.73%. The Innovative Intrusion Detection System is more effective than the Naive Bayes-based system since it uses Support Vector Machine (SVM). In any case, it’s crucial.