Paper Info
Title
A feature selection algorithm for intrusion detection system based on Moth Flame Optimization
Abstract
Machine Learning explores how machines can learn and develop from experience without being directly programmed. It is widely used for building intrusion detection systems. Supervised learning algorithms can "learn" from labeled dataset and then detect anomalies. It is capable of detecting new forms of intrusions, but it is vulnerable to false-positive warnings. Feature selection is critical in the development of machine learning models. Irrelevant features reduce model precision and lengthen the training time required to construct the model. In this paper, a wrapper feature selection algorithm for IDS is proposed. This algorithm uses the Moth Flame Optimization to utilize the selection process. The Moth Flame Optimization algorithm (MFO) is evolved from the lateral positioning and navigation mechanism of moths in nature. The proposed algorithm was evaluated using the CIC-2017 Dataset. The proposed algorithm is compared with a different wrapper technique: correlation-based feature selection using best-first search and compared with the case of taking all features without features selection. The results show that the number of features is reduced from 78 features to 4 features by using MFO, and after applying DT as a classifier, it gives a higher detection rate of 100% and an accuracy rate of 99.9% with a lower false alarm rate.
Year
DOI
Venue
2021
10.1109/ICIT52682.2021.9491690
2021 International Conference on Information Technology (ICIT)
Keywords
DocType
ISBN
Moth flame optimization,feature selection,decision tree,intrusion detection
Conference
978-1-6654-2871-2
Citations 
PageRank 
References 
0
0.34
0
Authors
2
Name
Order
Citations
PageRank
Arar Al Tawil100.34
Khair Eddin Sabri200.68