Marine Goal Optimizer Tuned Deep BiLSTM-Based Self-Configuring Intrusion Detection in Cloud

A Self-configuring intrusion detection system (IDS) present in the cloud monitors the suspicious activities affecting the user’s system and data by intruding on the stored resources. The traditional IDS environments analyze the available information for malicious detection, which makes clear that th...

Full description

Saved in:
Bibliographic Details
Published in:Journal of grid computing 2024-03, Vol.22 (1), p.24, Article 24
Main Authors: Bajpai, Sanchika Abhay, Patankar, Archana B.
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Classifiers
Computer Science
Datasets
Deep learning
Digital signatures
Distributed processing
Intrusion detection systems
Knowledge sharing
Management of Computing and Information Systems
Mathematical models
Parameter sensitivity
Processor Architectures
Support vector machines
Swarming
User Interfaces and Human Computer Interaction
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A Self-configuring intrusion detection system (IDS) present in the cloud monitors the suspicious activities affecting the user’s system and data by intruding on the stored resources. The traditional IDS environments analyze the available information for malicious detection, which makes clear that the manual analysis and attempts result in system failure. Thus, an automatic attack detection framework is needed for IDS, and this research proposes a marine goal optimizer-based deep BiLSTM classifier (MgLSTM model) for Self-configuring intrusion detection in the cloud. The significance of this research depends on the MgLSTM model for Self-configuring intrusion detection, which is developed by hybridizing the optimization with a deep classifier for effective detection. The algorithm follows the swarming, cooperative, and knowledge-sharing of marine predators, which renders an improved convergence rate for classifier parameters in the detection process. The analysis is performed by the datasets such as UNSW-NB 15as well as BoT-IoT datasets depending on the performance parameters, like specificity, sensitivity, and accuracy. The accuracy of the developed detection model is 99.26% with the UNSW-NB 15 database with reference to the training percentage and 99.20% with the BoT-IoT database with reference to the k-fold.
ISSN:1570-7873
1572-9184
DOI:10.1007/s10723-023-09728-0