Abstract: Intrusions that normally occur in computing systems are often meshed towards accessing, changing or damaging sensitive data or information. It is against this background, that various research has been carried out with the aim of solving detection and preventing such intrusive attacks.The similarity between the problem of computer security that is faced by Immune System (IS) can be shown by translating the language of immunology into computer security terms, Also the IS detects abuses of an implicitly specified policy, and responds to those abuses by counter-attacking the source of the abuse. However, Artificial Immune System (AIS) define the way the Human Immune System (HIS) responds to threats or attacks in the body. AIS and HIS are combined together by researchers to solve intrusion problems in Cybersecurity. The Negative Selection Algorithm (NSA) is an algorithm that divide the problem space into self and non-self which was used to build the model. In this study a model based on AIS concepts that will find a significant application in cybersecurity was developed and evaluated. The developed model called NNET NSA (Neural Network Negative Selection Algorithm) used the NSLKDDCup1999 dataset to test the model. The results from the developed model shows that the model NNET NSA achieved Receiver Operating Characteristics (ROC) showing 90% Area under the Curve (AUC) proportion of accuracy in detection of cyber-crime. The Error rate evaluation of NNET NSA classification of cyber-crime detection was the less by 0.05%, naïve Bayes by 0.16% and SVM by 0.22%. respectively on the R console.
Keywords: Artificial Immune System, Cybersecurity, Intrusion detection, Error Rate.
|
DOI:
10.17148/IJARCCE.2020.9603
