XGBoost Design by Multi-verse Optimiser: An Application for Network Intrusion Detection

Skup: ICMCSI 2022: Mobile Computing and Sustainable Informatics

Izdavač: Springer, Singapore

DOI: 10.1007/978-981-19-2069-1_1

Stranice: 1-16

Link: https://link.springer.com/chapter/10.1007/978-981-19-2069-1_1

This article presents the results of an experimental study, which aims to assess the efficiency of the performance of a novel multi-verse optimiser algorithm for the optimisation of parameters of a network intrusion detection system event classifier. The article gives an overview of computer network intrusion detection, outlines common issues faced by software solutions tackling this problem, and proposes using a machine learning algorithm to help solve some of these common issues. An XGBoost classification model with a multi-verse optimisation algorithm for adaptive search and optimisation is used to solve the network intrusion detection system event classifier hyper-parameter optimisation problem. Results of this experimental study are presented and discussed, the improvements compared to previous solutions are shown, and a possible direction of future work in this domain is given in the conclusion.
Ključne reči: Optimisation, Multi-verse optimiser, Intrusion detection, Computer networks, Parameter tuning, Machine learning