Original scientific paper
https://doi.org/10.32985/ijeces.15.5.1
Security Assessment Framework for IOT via Glove Optimized CNN-BiLSTM
Arun V
; Department of Computing Technologies, School of Computing, SRM Institute of Science and Technology, Kattankulathur, Chengalpattu 603203, India
*
Ramesh S
; Department of Computer Science Engineering, Krishnasamy College of Engineering Technology, Anand Nagar, Kumarapuram, Cuddalore, India
Carmel Sobia M
; Department of Electrical and Electronics Engineering, PSR Engineering College, Sivakasi, Tamil Nadu 626140, India
Geetha A
; Department of Electrical and Electronics Engineering, PSR Engineering College, Sivakasi, Tamil Nadu 626140, India
* Corresponding author.
Abstract
The Internet of Things (IoT) is a vast network of real, tangible objects or "things" that can communicate and share data with other systems and gadgets over the Internet. A vital component of assuring the secure and dependable operation of IoT systems and devices is IoT security. Attackers may use IoT devices to get unauthorized access, change functionality, or compromise the data that the device collects and transmits. The risks of IoT security breaches grow as more devices connect and exchange sensitive data. To check the vulnerability in Iot devices, a novel Blockchain enabled Iot based Security Assessment for intrusion (Block-ISAI) technique has been proposed. Bag of Words (BoW) technique is used for feature extraction of API documents which helps to make the document simpler. Blockchain technology is utilized for secure data storage and IoT device registration. In order to detect intrusion, a deep learning architecture is designed using the verified data The attack is either detected or not detected when the vulnerability is found using the GloVe-CNN-BiLSTM Model. If the vulnerability is detected then alerts will be given. Utilizing evaluation measures like accuracy, time efficiency, precision, F1 score, detection rate, recall, false alarm rate, usability and reliability the efficacy of the suggested ISAI technique has been assessed. By the comparison analysis, the proposed ISAI technique’s detection rate is 18.22%, 19.43%, and 3.13% higher than the existing HIDS, NIDS, and ML-DDoS techniques respectively. The accuracy of the proposed system is increased by 0.69%, 6.04%, and 36.15% as compared to the HIDS, NIDS, and ML-DDoS method using UNSW-NB 15 dataset and increases by 2.37%, 18.32%, and 5.95% using KDDCUP 19 dataset respectively.
Keywords
Internet of things (IoT); Security assessment; Vulnerabilities; Bag of words; deep learning;
Hrčak ID:
316781
URI
Publication date:
13.5.2024.
Visits: 405 *